【R-FCN】《R-FCN: Object Detection via Region-based Fully Convolutional Networks》

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/bryant_meng/article/details/81945236

NIPS-2016

目录

---

1 Motivation

Faster RCNN two stage ：fully convolutional subnetwork + RoI-wise subnetwork apply a costly per-region subnetwork hundreds of times.

为了进一步提速，作者用 fully convolutional with almost all computation shared on the entire image. 减少头部的计算时间，

但是呢，如果简单的把 fully connected（Alex、VGG）层换成 fully convolutional （ResNet、GoogleNet）会导致 inferior detection accuracy that does not match the network’s superior classification accuracy，这是因为 image-level classification task favors translation invariance 但是 object detection task needs localization representations that are translation-variant to an extent，会有冲突。

ResNet的做法是 creates a deeper RoI-wise subnetwork ，虽然提高了精度，但是速度有所下降 due to the unshared per-RoI computation.

本文的目的就要想办法解决这种冲突

2 Innovation

在faster RCNN 的基础上 propose position-sensitive score maps to address a dilemma between translation-invariance in image classification and translation-variance in object detection.

3 Advantages

• region-based detector is fully convolutional with almost all computation shared on the entire image
• 2.5-20× faster than the Faster R-CNN counterpart
• 83.6% mAP on the PASCAL VOC 2007，82.0% the 2012

4 Methods

4.1 Structures

• Backbone：ResNet101 = convolutional layers（100） + average pooling + 1000-class fc layer，作者只用100个convolutional layer来提取 feature map

• Position-sensitive score maps & Position-sensitive RoI pooling

最后一层 convolutional layer 的 channel 设计为 $k^2（C+1）$，C是类别数量

我们把score maps 上的 RoI 分成 k*k 个bins

ROI pooling 操作如下，

• $r_c(i, j)$ is the pooled response in the $(i,j)-th$ bin for the $c-th$ category
• $\Theta$ denotes all learnable parameters of the network
• $z_{i,j,c}$ is one of score map out of the $k^2(C+1)$ score maps
• $(x_0,y_0)$ denotes the top-left corner of an RoI
• $n$ is the number of pixels in the bin

需要注意的是，pooling 后的第 i,j,c-th bin 是由 第 i,j,c-th score map对应的第 i.j bin 累加得到 ，也即是各类中每个 score maps 只 pooling 某一个bins，比如白色的 score maps 只对右下角部分的像素进行操作。

作者 address bbox regression in a similar way, feature map 通过 $4k^2$ channel 的 conv 产生 $4k^2$ channel 的 score maps，然后 producing a $4k^2$-d vector for each RoI，接着 vote 成 4-d vector，表示 $（t_x，t_y，t_w，t_h）$，这种情况是 class-agnostic，对于 class-specific 来说， $4k^2*C$ 即可。

4.2 train

• End to End
• Loss function，和 Fast RCNN 一样 【Fast RCNN】《Fast-RCNN》
  
  分类 Loss 和 回归 Loss 通过一个参数 λ （论文中为1）加权平均，$L_{reg}$ 同样采用 L1 Smooth。
• Positive：IoU at least 0.5，Negative otherwise
• OHEM
• Single scales training：shorter side of image is 600 pixels
• 4-steps 训练，和 Faster RCNN 一样 【Faster RCNN】《Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks》，只不过是 RPN 和 R-FCN的交替训练

4.3 inference

• 300 RoI
• non-maximum suppression：0.3

4.4 À trous and stride

ResNet 101 中， stage 1-4 unchanged，stage 5 中，把第一个 stride = 2 的conv 变成1，然后用 À trous to compensate for the reduced stride

这样 ResNet stride from 32 to 16，提高了 score maps 的分辨率

提升效果如下，2.6 个 点

4.5 Visualization

如果每个bins 的响应都很高，vote 后的 score 会很高，反之，则 low

5 Experiments

5.1 Experiments on PASCAL VOC

5.1.1 验证结构的效果

• training ：07+12
• test：07

• Standard Faster RCNN：76.4% mAP，RoI pooling 插入 stage4 和 stage5之间（见文章最后的 Faster RCNN-RES101 结构图）
• naive Faster RCNN：RoI pooling after conv5，an inexpensive 21-class fc layer is evaluated on each RoI，为了对比的公平，用了 À trous trick
• class-specific RPN：RPN 的二分类（object or not）换成了 21 类，也用了À trous trick
• R-FCN（without position-sensitivity）k = 1，相当于 global pooling within each RoI

5.1.2 与Faster RCNN pk，速度和精度，k*k = 7*7

• 单技能释放PK
  mining from a larger set of candidates has no benefit，所以作者采用 RoI = 300 （both training and inference）
  

  

• 组合拳PK
  

  

Faster R-CNN +++ 就是 ResNet 论文里面的 Box refinement + Global context + Multi-scale testing

5.1.3 Impact of Depth

5.1.4 Impact of Region Proposals

5.2 Experiments on MS COCO

R-FCN vs Faster RCNN 不打组合拳是 R-FCN厉害，组合后 Faster RCNN 厉害，但是 R-FCN 会快很多

---

Faster RCNN-RES101 结构图

train.prototxt

# Enter your network definition here.
# Use Shift+Enter to update the visualization.
name: "ResNet-50"
layer {
  name: 'input-data'
  type: 'Python'
  top: 'data'
  top: 'im_info'
  top: 'gt_boxes'
  python_param {
    module: 'roi_data_layer.layer'
    layer: 'RoIDataLayer'
    param_str: "'num_classes': 21"
  }
}

# ------------------------ conv1 -----------------------------
layer {
    bottom: "data"
    top: "conv1"
    name: "conv1"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 7
        pad: 3
        stride: 2
    }
    param {
        lr_mult: 0.0
    }
    param {
        lr_mult: 0.0
    }

}

layer {
    bottom: "conv1"
    top: "conv1"
    name: "bn_conv1"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "conv1"
    top: "conv1"
    name: "scale_conv1"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "conv1"
    top: "conv1"
    name: "conv1_relu"
    type: "ReLU"
}

layer {
    bottom: "conv1"
    top: "pool1"
    name: "pool1"
    type: "Pooling"
    pooling_param {
        kernel_size: 3
        stride: 2
        pool: MAX
    }
}

layer {
    bottom: "pool1"
    top: "res2a_branch1"
    name: "res2a_branch1"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch1"
    top: "res2a_branch1"
    name: "bn2a_branch1"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch1"
    top: "res2a_branch1"
    name: "scale2a_branch1"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "pool1"
    top: "res2a_branch2a"
    name: "res2a_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2a"
    top: "res2a_branch2a"
    name: "bn2a_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2a"
    top: "res2a_branch2a"
    name: "scale2a_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2a"
    top: "res2a_branch2a"
    name: "res2a_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res2a_branch2a"
    top: "res2a_branch2b"
    name: "res2a_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2b"
    top: "res2a_branch2b"
    name: "bn2a_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2b"
    top: "res2a_branch2b"
    name: "scale2a_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2b"
    top: "res2a_branch2b"
    name: "res2a_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res2a_branch2b"
    top: "res2a_branch2c"
    name: "res2a_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2c"
    top: "res2a_branch2c"
    name: "bn2a_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch2c"
    top: "res2a_branch2c"
    name: "scale2a_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a_branch1"
    bottom: "res2a_branch2c"
    top: "res2a"
    name: "res2a"
    type: "Eltwise"
}

layer {
    bottom: "res2a"
    top: "res2a"
    name: "res2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res2a"
    top: "res2b_branch2a"
    name: "res2b_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2a"
    top: "res2b_branch2a"
    name: "bn2b_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2a"
    top: "res2b_branch2a"
    name: "scale2b_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2a"
    top: "res2b_branch2a"
    name: "res2b_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res2b_branch2a"
    top: "res2b_branch2b"
    name: "res2b_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2b"
    top: "res2b_branch2b"
    name: "bn2b_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2b"
    top: "res2b_branch2b"
    name: "scale2b_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2b"
    top: "res2b_branch2b"
    name: "res2b_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res2b_branch2b"
    top: "res2b_branch2c"
    name: "res2b_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2c"
    top: "res2b_branch2c"
    name: "bn2b_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b_branch2c"
    top: "res2b_branch2c"
    name: "scale2b_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2a"
    bottom: "res2b_branch2c"
    top: "res2b"
    name: "res2b"
    type: "Eltwise"
}

layer {
    bottom: "res2b"
    top: "res2b"
    name: "res2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res2b"
    top: "res2c_branch2a"
    name: "res2c_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2a"
    top: "res2c_branch2a"
    name: "bn2c_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2a"
    top: "res2c_branch2a"
    name: "scale2c_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2a"
    top: "res2c_branch2a"
    name: "res2c_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res2c_branch2a"
    top: "res2c_branch2b"
    name: "res2c_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 64
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2b"
    top: "res2c_branch2b"
    name: "bn2c_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2b"
    top: "res2c_branch2b"
    name: "scale2c_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2b"
    top: "res2c_branch2b"
    name: "res2c_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res2c_branch2b"
    top: "res2c_branch2c"
    name: "res2c_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2c"
    top: "res2c_branch2c"
    name: "bn2c_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c_branch2c"
    top: "res2c_branch2c"
    name: "scale2c_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2b"
    bottom: "res2c_branch2c"
    top: "res2c"
    name: "res2c"
    type: "Eltwise"
}

layer {
    bottom: "res2c"
    top: "res2c"
    name: "res2c_relu"
    type: "ReLU"
}

layer {
    bottom: "res2c"
    top: "res3a_branch1"
    name: "res3a_branch1"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 2
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3a_branch1"
    top: "res3a_branch1"
    name: "bn3a_branch1"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch1"
    top: "res3a_branch1"
    name: "scale3a_branch1"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res2c"
    top: "res3a_branch2a"
    name: "res3a_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 1
        pad: 0
        stride: 2
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3a_branch2a"
    top: "res3a_branch2a"
    name: "bn3a_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch2a"
    top: "res3a_branch2a"
    name: "scale3a_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch2a"
    top: "res3a_branch2a"
    name: "res3a_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res3a_branch2a"
    top: "res3a_branch2b"
    name: "res3a_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3a_branch2b"
    top: "res3a_branch2b"
    name: "bn3a_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch2b"
    top: "res3a_branch2b"
    name: "scale3a_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch2b"
    top: "res3a_branch2b"
    name: "res3a_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res3a_branch2b"
    top: "res3a_branch2c"
    name: "res3a_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3a_branch2c"
    top: "res3a_branch2c"
    name: "bn3a_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch2c"
    top: "res3a_branch2c"
    name: "scale3a_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a_branch1"
    bottom: "res3a_branch2c"
    top: "res3a"
    name: "res3a"
    type: "Eltwise"
}

layer {
    bottom: "res3a"
    top: "res3a"
    name: "res3a_relu"
    type: "ReLU"
}

layer {
    bottom: "res3a"
    top: "res3b_branch2a"
    name: "res3b_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3b_branch2a"
    top: "res3b_branch2a"
    name: "bn3b_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b_branch2a"
    top: "res3b_branch2a"
    name: "scale3b_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b_branch2a"
    top: "res3b_branch2a"
    name: "res3b_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res3b_branch2a"
    top: "res3b_branch2b"
    name: "res3b_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3b_branch2b"
    top: "res3b_branch2b"
    name: "bn3b_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b_branch2b"
    top: "res3b_branch2b"
    name: "scale3b_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b_branch2b"
    top: "res3b_branch2b"
    name: "res3b_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res3b_branch2b"
    top: "res3b_branch2c"
    name: "res3b_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3b_branch2c"
    top: "res3b_branch2c"
    name: "bn3b_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b_branch2c"
    top: "res3b_branch2c"
    name: "scale3b_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3a"
    bottom: "res3b_branch2c"
    top: "res3b"
    name: "res3b"
    type: "Eltwise"
}

layer {
    bottom: "res3b"
    top: "res3b"
    name: "res3b_relu"
    type: "ReLU"
}

layer {
    bottom: "res3b"
    top: "res3c_branch2a"
    name: "res3c_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3c_branch2a"
    top: "res3c_branch2a"
    name: "bn3c_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c_branch2a"
    top: "res3c_branch2a"
    name: "scale3c_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c_branch2a"
    top: "res3c_branch2a"
    name: "res3c_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res3c_branch2a"
    top: "res3c_branch2b"
    name: "res3c_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3c_branch2b"
    top: "res3c_branch2b"
    name: "bn3c_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c_branch2b"
    top: "res3c_branch2b"
    name: "scale3c_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c_branch2b"
    top: "res3c_branch2b"
    name: "res3c_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res3c_branch2b"
    top: "res3c_branch2c"
    name: "res3c_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3c_branch2c"
    top: "res3c_branch2c"
    name: "bn3c_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c_branch2c"
    top: "res3c_branch2c"
    name: "scale3c_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3b"
    bottom: "res3c_branch2c"
    top: "res3c"
    name: "res3c"
    type: "Eltwise"
}

layer {
    bottom: "res3c"
    top: "res3c"
    name: "res3c_relu"
    type: "ReLU"
}

layer {
    bottom: "res3c"
    top: "res3d_branch2a"
    name: "res3d_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3d_branch2a"
    top: "res3d_branch2a"
    name: "bn3d_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d_branch2a"
    top: "res3d_branch2a"
    name: "scale3d_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d_branch2a"
    top: "res3d_branch2a"
    name: "res3d_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res3d_branch2a"
    top: "res3d_branch2b"
    name: "res3d_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 128
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3d_branch2b"
    top: "res3d_branch2b"
    name: "bn3d_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d_branch2b"
    top: "res3d_branch2b"
    name: "scale3d_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d_branch2b"
    top: "res3d_branch2b"
    name: "res3d_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res3d_branch2b"
    top: "res3d_branch2c"
    name: "res3d_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res3d_branch2c"
    top: "res3d_branch2c"
    name: "bn3d_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d_branch2c"
    top: "res3d_branch2c"
    name: "scale3d_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3c"
    bottom: "res3d_branch2c"
    top: "res3d"
    name: "res3d"
    type: "Eltwise"
}

layer {
    bottom: "res3d"
    top: "res3d"
    name: "res3d_relu"
    type: "ReLU"
}

layer {
    bottom: "res3d"
    top: "res4a_branch1"
    name: "res4a_branch1"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 2
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4a_branch1"
    top: "res4a_branch1"
    name: "bn4a_branch1"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch1"
    top: "res4a_branch1"
    name: "scale4a_branch1"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res3d"
    top: "res4a_branch2a"
    name: "res4a_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 2
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4a_branch2a"
    top: "res4a_branch2a"
    name: "bn4a_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch2a"
    top: "res4a_branch2a"
    name: "scale4a_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch2a"
    top: "res4a_branch2a"
    name: "res4a_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4a_branch2a"
    top: "res4a_branch2b"
    name: "res4a_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4a_branch2b"
    top: "res4a_branch2b"
    name: "bn4a_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch2b"
    top: "res4a_branch2b"
    name: "scale4a_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch2b"
    top: "res4a_branch2b"
    name: "res4a_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4a_branch2b"
    top: "res4a_branch2c"
    name: "res4a_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4a_branch2c"
    top: "res4a_branch2c"
    name: "bn4a_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch2c"
    top: "res4a_branch2c"
    name: "scale4a_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a_branch1"
    bottom: "res4a_branch2c"
    top: "res4a"
    name: "res4a"
    type: "Eltwise"
}

layer {
    bottom: "res4a"
    top: "res4a"
    name: "res4a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4a"
    top: "res4b_branch2a"
    name: "res4b_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4b_branch2a"
    top: "res4b_branch2a"
    name: "bn4b_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b_branch2a"
    top: "res4b_branch2a"
    name: "scale4b_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b_branch2a"
    top: "res4b_branch2a"
    name: "res4b_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4b_branch2a"
    top: "res4b_branch2b"
    name: "res4b_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4b_branch2b"
    top: "res4b_branch2b"
    name: "bn4b_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b_branch2b"
    top: "res4b_branch2b"
    name: "scale4b_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b_branch2b"
    top: "res4b_branch2b"
    name: "res4b_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4b_branch2b"
    top: "res4b_branch2c"
    name: "res4b_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4b_branch2c"
    top: "res4b_branch2c"
    name: "bn4b_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b_branch2c"
    top: "res4b_branch2c"
    name: "scale4b_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4a"
    bottom: "res4b_branch2c"
    top: "res4b"
    name: "res4b"
    type: "Eltwise"
}

layer {
    bottom: "res4b"
    top: "res4b"
    name: "res4b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4b"
    top: "res4c_branch2a"
    name: "res4c_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4c_branch2a"
    top: "res4c_branch2a"
    name: "bn4c_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c_branch2a"
    top: "res4c_branch2a"
    name: "scale4c_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c_branch2a"
    top: "res4c_branch2a"
    name: "res4c_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4c_branch2a"
    top: "res4c_branch2b"
    name: "res4c_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4c_branch2b"
    top: "res4c_branch2b"
    name: "bn4c_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c_branch2b"
    top: "res4c_branch2b"
    name: "scale4c_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c_branch2b"
    top: "res4c_branch2b"
    name: "res4c_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4c_branch2b"
    top: "res4c_branch2c"
    name: "res4c_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4c_branch2c"
    top: "res4c_branch2c"
    name: "bn4c_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c_branch2c"
    top: "res4c_branch2c"
    name: "scale4c_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4b"
    bottom: "res4c_branch2c"
    top: "res4c"
    name: "res4c"
    type: "Eltwise"
}

layer {
    bottom: "res4c"
    top: "res4c"
    name: "res4c_relu"
    type: "ReLU"
}

layer {
    bottom: "res4c"
    top: "res4d_branch2a"
    name: "res4d_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4d_branch2a"
    top: "res4d_branch2a"
    name: "bn4d_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d_branch2a"
    top: "res4d_branch2a"
    name: "scale4d_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d_branch2a"
    top: "res4d_branch2a"
    name: "res4d_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4d_branch2a"
    top: "res4d_branch2b"
    name: "res4d_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4d_branch2b"
    top: "res4d_branch2b"
    name: "bn4d_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d_branch2b"
    top: "res4d_branch2b"
    name: "scale4d_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d_branch2b"
    top: "res4d_branch2b"
    name: "res4d_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4d_branch2b"
    top: "res4d_branch2c"
    name: "res4d_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4d_branch2c"
    top: "res4d_branch2c"
    name: "bn4d_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d_branch2c"
    top: "res4d_branch2c"
    name: "scale4d_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4c"
    bottom: "res4d_branch2c"
    top: "res4d"
    name: "res4d"
    type: "Eltwise"
}

layer {
    bottom: "res4d"
    top: "res4d"
    name: "res4d_relu"
    type: "ReLU"
}

layer {
    bottom: "res4d"
    top: "res4e_branch2a"
    name: "res4e_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4e_branch2a"
    top: "res4e_branch2a"
    name: "bn4e_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e_branch2a"
    top: "res4e_branch2a"
    name: "scale4e_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e_branch2a"
    top: "res4e_branch2a"
    name: "res4e_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4e_branch2a"
    top: "res4e_branch2b"
    name: "res4e_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4e_branch2b"
    top: "res4e_branch2b"
    name: "bn4e_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e_branch2b"
    top: "res4e_branch2b"
    name: "scale4e_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e_branch2b"
    top: "res4e_branch2b"
    name: "res4e_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4e_branch2b"
    top: "res4e_branch2c"
    name: "res4e_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4e_branch2c"
    top: "res4e_branch2c"
    name: "bn4e_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e_branch2c"
    top: "res4e_branch2c"
    name: "scale4e_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4d"
    bottom: "res4e_branch2c"
    top: "res4e"
    name: "res4e"
    type: "Eltwise"
}

layer {
    bottom: "res4e"
    top: "res4e"
    name: "res4e_relu"
    type: "ReLU"
}

layer {
    bottom: "res4e"
    top: "res4f_branch2a"
    name: "res4f_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4f_branch2a"
    top: "res4f_branch2a"
    name: "bn4f_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f_branch2a"
    top: "res4f_branch2a"
    name: "scale4f_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f_branch2a"
    top: "res4f_branch2a"
    name: "res4f_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res4f_branch2a"
    top: "res4f_branch2b"
    name: "res4f_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 256
        kernel_size: 3
        pad: 1
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4f_branch2b"
    top: "res4f_branch2b"
    name: "bn4f_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f_branch2b"
    top: "res4f_branch2b"
    name: "scale4f_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f_branch2b"
    top: "res4f_branch2b"
    name: "res4f_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res4f_branch2b"
    top: "res4f_branch2c"
    name: "res4f_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res4f_branch2c"
    top: "res4f_branch2c"
    name: "bn4f_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f_branch2c"
    top: "res4f_branch2c"
    name: "scale4f_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4e"
    bottom: "res4f_branch2c"
    top: "res4f"
    name: "res4f"
    type: "Eltwise"
}

layer {
    bottom: "res4f"
    top: "res4f"
    name: "res4f_relu"
    type: "ReLU"
}

layer {
    bottom: "res4f"
    top: "res5a_branch1"
    name: "res5a_branch1"
    type: "Convolution"
    convolution_param {
        num_output: 2048
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5a_branch1"
    top: "res5a_branch1"
    name: "bn5a_branch1"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch1"
    top: "res5a_branch1"
    name: "scale5a_branch1"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res4f"
    top: "res5a_branch2a"
    name: "res5a_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5a_branch2a"
    top: "res5a_branch2a"
    name: "bn5a_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch2a"
    top: "res5a_branch2a"
    name: "scale5a_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch2a"
    top: "res5a_branch2a"
    name: "res5a_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res5a_branch2a"
    top: "res5a_branch2b"
    name: "res5a_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 3
        dilation: 2
        pad: 2
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5a_branch2b"
    top: "res5a_branch2b"
    name: "bn5a_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch2b"
    top: "res5a_branch2b"
    name: "scale5a_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch2b"
    top: "res5a_branch2b"
    name: "res5a_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res5a_branch2b"
    top: "res5a_branch2c"
    name: "res5a_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 2048
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5a_branch2c"
    top: "res5a_branch2c"
    name: "bn5a_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch2c"
    top: "res5a_branch2c"
    name: "scale5a_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a_branch1"
    bottom: "res5a_branch2c"
    top: "res5a"
    name: "res5a"
    type: "Eltwise"
}

layer {
    bottom: "res5a"
    top: "res5a"
    name: "res5a_relu"
    type: "ReLU"
}

layer {
    bottom: "res5a"
    top: "res5b_branch2a"
    name: "res5b_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5b_branch2a"
    top: "res5b_branch2a"
    name: "bn5b_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b_branch2a"
    top: "res5b_branch2a"
    name: "scale5b_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b_branch2a"
    top: "res5b_branch2a"
    name: "res5b_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res5b_branch2a"
    top: "res5b_branch2b"
    name: "res5b_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 3
        dilation: 2
        pad: 2
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5b_branch2b"
    top: "res5b_branch2b"
    name: "bn5b_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b_branch2b"
    top: "res5b_branch2b"
    name: "scale5b_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b_branch2b"
    top: "res5b_branch2b"
    name: "res5b_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res5b_branch2b"
    top: "res5b_branch2c"
    name: "res5b_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 2048
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5b_branch2c"
    top: "res5b_branch2c"
    name: "bn5b_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b_branch2c"
    top: "res5b_branch2c"
    name: "scale5b_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5a"
    bottom: "res5b_branch2c"
    top: "res5b"
    name: "res5b"
    type: "Eltwise"
}

layer {
    bottom: "res5b"
    top: "res5b"
    name: "res5b_relu"
    type: "ReLU"
}

layer {
    bottom: "res5b"
    top: "res5c_branch2a"
    name: "res5c_branch2a"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5c_branch2a"
    top: "res5c_branch2a"
    name: "bn5c_branch2a"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5c_branch2a"
    top: "res5c_branch2a"
    name: "scale5c_branch2a"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5c_branch2a"
    top: "res5c_branch2a"
    name: "res5c_branch2a_relu"
    type: "ReLU"
}

layer {
    bottom: "res5c_branch2a"
    top: "res5c_branch2b"
    name: "res5c_branch2b"
    type: "Convolution"
    convolution_param {
        num_output: 512
        kernel_size: 3
        dilation: 2
        pad: 2
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5c_branch2b"
    top: "res5c_branch2b"
    name: "bn5c_branch2b"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5c_branch2b"
    top: "res5c_branch2b"
    name: "scale5c_branch2b"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5c_branch2b"
    top: "res5c_branch2b"
    name: "res5c_branch2b_relu"
    type: "ReLU"
}

layer {
    bottom: "res5c_branch2b"
    top: "res5c_branch2c"
    name: "res5c_branch2c"
    type: "Convolution"
    convolution_param {
        num_output: 2048
        kernel_size: 1
        pad: 0
        stride: 1
        bias_term: false
    }
    param {
        lr_mult: 1.0
    }
}

layer {
    bottom: "res5c_branch2c"
    top: "res5c_branch2c"
    name: "bn5c_branch2c"
    type: "BatchNorm"
    batch_norm_param {
        use_global_stats: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5c_branch2c"
    top: "res5c_branch2c"
    name: "scale5c_branch2c"
    type: "Scale"
    scale_param {
        bias_term: true
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
    param {
        lr_mult: 0.0
        decay_mult: 0.0
    }
}

layer {
    bottom: "res5b"
    bottom: "res5c_branch2c"
    top: "res5c"
    name: "res5c"
    type: "Eltwise"
}

layer {
    bottom: "res5c"
    top: "res5c"
    name: "res5c_relu"
    type: "ReLU"
}


#========= RPN ============

layer {
  name: "rpn_conv/3x3"
  type: "Convolution"
  bottom: "res4f"
  top: "rpn/output"
  param { lr_mult: 1.0 }
  param { lr_mult: 2.0 }
  convolution_param {
    num_output: 512
    kernel_size: 3 pad: 1 stride: 1
    weight_filler { type: "gaussian" std: 0.01 }
    bias_filler { type: "constant" value: 0 }
  }
}
layer {
  name: "rpn_relu/3x3"
  type: "ReLU"
  bottom: "rpn/output"
  top: "rpn/output"
}

layer {
  name: "rpn_cls_score"
  type: "Convolution"
  bottom: "rpn/output"
  top: "rpn_cls_score"
  param { lr_mult: 1.0 }
  param { lr_mult: 2.0 }
  convolution_param {
    num_output: 18   # 2(bg/fg) * 9(anchors)
    kernel_size: 1 pad: 0 stride: 1
    weight_filler { type: "gaussian" std: 0.01 }
    bias_filler { type: "constant" value: 0 }
  }
}

layer {
  name: "rpn_bbox_pred"
  type: "Convolution"
  bottom: "rpn/output"
  top: "rpn_bbox_pred"
  param { lr_mult: 1.0 }
  param { lr_mult: 2.0 }
  convolution_param {
    num_output: 36   # 4 * 9(anchors)
    kernel_size: 1 pad: 0 stride: 1
    weight_filler { type: "gaussian" std: 0.01 }
    bias_filler { type: "constant" value: 0 }
  }
}

layer {
   bottom: "rpn_cls_score"
   top: "rpn_cls_score_reshape"
   name: "rpn_cls_score_reshape"
   type: "Reshape"
   reshape_param { shape { dim: 0 dim: 2 dim: -1 dim: 0 } }
}

layer {
  name: 'rpn-data'
  type: 'Python'
  bottom: 'rpn_cls_score'
  bottom: 'gt_boxes'
  bottom: 'im_info'
  bottom: 'data'
  top: 'rpn_labels'
  top: 'rpn_bbox_targets'
  top: 'rpn_bbox_inside_weights'
  top: 'rpn_bbox_outside_weights'
  python_param {
    module: 'rpn.anchor_target_layer'
    layer: 'AnchorTargetLayer'
    param_str: "'feat_stride': 16"
  }
}

layer {
  name: "rpn_loss_cls"
  type: "SoftmaxWithLoss"
  bottom: "rpn_cls_score_reshape"
  bottom: "rpn_labels"
  propagate_down: 1
  propagate_down: 0
  top: "rpn_cls_loss"
  loss_weight: 1
  loss_param {
    ignore_label: -1
    normalize: true
  }
}

layer {
  name: "rpn_loss_bbox"
  type: "SmoothL1Loss"
  bottom: "rpn_bbox_pred"
  bottom: "rpn_bbox_targets"
  bottom: 'rpn_bbox_inside_weights'
  bottom: 'rpn_bbox_outside_weights'
  top: "rpn_loss_bbox"
  loss_weight: 1
  smooth_l1_loss_param { sigma: 3.0 }
}

#========= RoI Proposal ============

layer {
  name: "rpn_cls_prob"
  type: "Softmax"
  bottom: "rpn_cls_score_reshape"
  top: "rpn_cls_prob"
}

layer {
  name: 'rpn_cls_prob_reshape'
  type: 'Reshape'
  bottom: 'rpn_cls_prob'
  top: 'rpn_cls_prob_reshape'
  reshape_param { shape { dim: 0 dim: 18 dim: -1 dim: 0 } }
}

layer {
  name: 'proposal'
  type: 'Python'
  bottom: 'rpn_cls_prob_reshape'
  bottom: 'rpn_bbox_pred'
  bottom: 'im_info'
  top: 'rpn_rois'
#  top: 'rpn_scores'
  python_param {
    module: 'rpn.proposal_layer'
    layer: 'ProposalLayer'
    param_str: "'feat_stride': 16"
  }
}

#layer {
#  name: 'debug-data'
#  type: 'Python'
#  bottom: 'data'
#  bottom: 'rpn_rois'
#  bottom: 'rpn_scores'
#  python_param {
#    module: 'rpn.debug_layer'
#    layer: 'RPNDebugLayer'
#  }
#}

layer {
  name: 'roi-data'
  type: 'Python'
  bottom: 'rpn_rois'
  bottom: 'gt_boxes'
  top: 'rois'
  top: 'labels'
  top: 'bbox_targets'
  top: 'bbox_inside_weights'
  top: 'bbox_outside_weights'
  python_param {
    module: 'rpn.proposal_target_layer'
    layer: 'ProposalTargetLayer'
    param_str: "'num_classes': 2"
  }
}

#----------------------new conv layer------------------
layer {
    bottom: "res5c"
    top: "conv_new_1"
    name: "conv_new_1"
    type: "Convolution"
    convolution_param {
        num_output: 1024
        kernel_size: 1
        pad: 0
        weight_filler {
            type: "gaussian"
            std: 0.01
        }
        bias_filler {
            type: "constant"
            value: 0
        }
    }
    param {
        lr_mult: 1.0
    }
    param {
        lr_mult: 2.0
    }
}

layer {
    bottom: "conv_new_1"
    top: "conv_new_1"
    name: "conv_new_1_relu"
    type: "ReLU"
}

layer {
    bottom: "conv_new_1"
    top: "rfcn_cls"
    name: "rfcn_cls"
    type: "Convolution"
    convolution_param {
        num_output: 1029 #21*(7^2) cls_num*(score_maps_size^2)
        kernel_size: 1
        pad: 0
        weight_filler {
            type: "gaussian"
            std: 0.01
        }
        bias_filler {
            type: "constant"
            value: 0
        }
    }
    param {
        lr_mult: 1.0
    }
    param {
        lr_mult: 2.0
    }
}
layer {
    bottom: "conv_new_1"
    top: "rfcn_bbox"
    name: "rfcn_bbox"
    type: "Convolution"
    convolution_param {
        num_output: 392 #2*4*(7^2) (bg/fg)*(dx, dy, dw, dh)*(score_maps_size^2)
        kernel_size: 1
        pad: 0
        weight_filler {
            type: "gaussian"
            std: 0.01
        }
        bias_filler {
            type: "constant"
            value: 0
        }
    }
    param {
        lr_mult: 1.0
    }
    param {
        lr_mult: 2.0
    }
}

#--------------position sensitive RoI pooling--------------
layer {
    bottom: "rfcn_cls"
    bottom: "rois"
    top: "psroipooled_cls_rois"
    name: "psroipooled_cls_rois"
    type: "PSROIPooling"
    psroi_pooling_param {
        spatial_scale: 0.0625
        output_dim: 21
        group_size: 7
    }
}

layer {
    bottom: "psroipooled_cls_rois"
    top: "cls_score"
    name: "ave_cls_score_rois"
    type: "Pooling"
    pooling_param {
        pool: AVE
        kernel_size: 7
        stride: 7
    }
}


layer {
    bottom: "rfcn_bbox"
    bottom: "rois"
    top: "psroipooled_loc_rois"
    name: "psroipooled_loc_rois"
    type: "PSROIPooling"
    psroi_pooling_param {
        spatial_scale: 0.0625
        output_dim: 8
        group_size: 7
    }
}

layer {
    bottom: "psroipooled_loc_rois"
    top: "bbox_pred"
    name: "ave_bbox_pred_rois"
    type: "Pooling"
    pooling_param {
        pool: AVE
        kernel_size: 7
        stride: 7
    }
}


#-----------------------output------------------------
layer {
   name: "loss"
   type: "SoftmaxWithLoss"
   bottom: "cls_score"
   bottom: "labels"
   top: "loss_cls"
   loss_weight: 1
   propagate_down: true
   propagate_down: false
}

layer {
   name: "accuarcy"
   type: "Accuracy"
   bottom: "cls_score"
   bottom: "labels"
   top: "accuarcy"
   #include: { phase: TEST }
   propagate_down: false
   propagate_down: false
}

layer {
   name: "loss_bbox"
   type: "SmoothL1LossOHEM"
   bottom: "bbox_pred"
   bottom: "bbox_targets"
   bottom: 'bbox_inside_weights'
   top: "loss_bbox"
   loss_weight: 1
   loss_param {
        normalization: PRE_FIXED
        pre_fixed_normalizer: 128
   }
   propagate_down: true
   propagate_down: false
   propagate_down: false
}

layer {
    name: "silence"
    type: "Silence"
    bottom: "bbox_outside_weights"
}

参考

【1】Faster R-CNN改进篇（二）： RFCN ● RON
【2】RFCN论文笔记