【目标检测系列：五】2018 CVPR IoU-Net 论文阅读解析总结

2018 CVPR

建议先自己看一遍论文，然后再看下面的总结

IoU-Net

解决问题： nms 过程中，是挑选分类置信度最大的值的框，但是它不一定框的准

Two drawbacks in object localization

Training

Training the IoU detector

Inference

first apply bounding box regression for the initial coordinates
IoU-guide NMS
on all detected bounding boxes
refine using optimization-based algorithm
100 bounding boxes with highest classification confidence

IoU predictor

aim
- takes features from the FPN
- estimates the localization accuracy (IoU) for each bounding box
data generation
- generate candidate bounding box set
  generate bounding boxes and labels for training the IoU-Net : augmenting the ground-truth,instead of taking proposals from RPNs
  for all ground-truth bounding box in training set , manually transform them with a set of randomized parameters
- remove the bounding box having an IoU < 0.5 with the matched ground-truth
feature
- extracted from the output of FPN with the proposed PrRoI-Pooling layers
- then fed into a two-layer feedforward network for the IoU prediction
use class-aware IoU predictors

IoU-guided NMS

use the predicted IoU instead of the classification confidence as the ranking keyword for bounding boxes.
to determine the classification scores
- select the box having the highest IoU with a ground-truth
- eliminate all other boxes having an overlap greater than threshold nms
- for a group of bounding boxes matching the same ground-truth, we take the most confident prediction for the class label.
  highest IoU 的框的分类置信度是其和他匹配同一gt的并大于阈值被滤掉的框的分类置信度的最大值
Algorithm
- ① 从bounding box集合 B 中依次选取预估IOU（localization confidence）最高的bounding box（记为 $b_m$ ）
- ② 将与其IOU高于一定阈值的bounding box一个个选出来，并将这些bounding box（包括最开始选的 $b_m$ ）的最高classification confidence记为 $s$
- ③ 将 $b_m,s)$ 二元组记录到集合 D 中 (本质是 bounding box和cls conf的重新分配)

Optimization-based bounding box refinement

Algorithm
- 对于检测到的bounding box，利用 PrPool 提取内部特征并算出 IOUnet 预测的IOU，记其梯度为grad，这个IOU记为PrevScore
- 然后更新bounding box
- 更新之后重新进行IOU预测结果为NewScore
- 如果 prevscore 和 newscore 相差小于一个early-stop阈值或者 newscore 比 prevscore 低于一个“定位退化容忍度”，则认为该bounding box更新完毕。

PrPool