Visualization Pytorch designated layer (Udacity)

import cv2
import matplotlib.pyplot as plt
%matplotlib inline

# TODO: Feel free to try out your own images here by changing img_path
# to a file path to another image on your computer!
img_path = 'images/udacity_sdc.png'

# load color image 
bgr_img = cv2.imread(img_path)
# convert to grayscale
gray_img = cv2.cvtColor(bgr_img, cv2.COLOR_BGR2GRAY)

# normalize, rescale entries to lie in [0,1]
gray_img = gray_img.astype("float32")/255

# plot image
plt.imshow(gray_img, cmap='gray')
plt.show()

Display image

Define the filter, and visualize

import numpy as np

## TODO: Feel free to modify the numbers here, to try out another filter!
filter_vals = np.array([[-1, -1, 1, 1], [-1, -1, 1, 1], [-1, -1, 1, 1], [-1, -1, 1, 1]])

print('Filter shape: ', filter_vals.shape)

Filter shape:  (4, 4)

# Defining four different filters, 
# all of which are linear combinations of the `filter_vals` defined above

# define four filters
filter_1 = filter_vals
filter_2 = -filter_1
filter_3 = filter_1.T
filter_4 = -filter_3
filters = np.array([filter_1, filter_2, filter_3, filter_4])

# For an example, print out the values of filter 1
print('Filter 1: \n', filter_1)

 Filter 1: [[-1 -1 1 1] [-1 -1 1 1] [-1 -1 1 1] [-1 -1 1 1]]

Defined cell layer and a convolution layer ¶

The convolution layer initialized to contain all the filters you've created. Then add a layer of the largest pools ( related documents please click here to view ), the kernel size (4x4), so you can see, after this step, the image resolution has been reduced!

import torch
import torch.nn as nn
import torch.nn.functional as F

    
# define a neural network with a convolutional layer with four filters
# AND a pooling layer of size (4, 4)
class Net(nn.Module):
    
    def __init__(self, weight):
        super(Net, self).__init__()
        # initializes the weights of the convolutional layer to be the weights of the 4 defined filters
        k_height, k_width = weight.shape[2:]
        # assumes there are 4 grayscale filters
        self.conv = nn.Conv2d(1, 4, kernel_size=(k_height, k_width), bias=False)
        self.conv.weight = torch.nn.Parameter(weight)
        # define a pooling layer
        self.pool = nn.MaxPool2d(4, 4)

    def forward(self, x):
        # calculates the output of a convolutional layer
        # pre- and post-activation
        conv_x = self.conv(x)
        activated_x = F.relu(conv_x)
        
        # applies pooling layer
        pooled_x = self.pool(activated_x)
        
        # returns all layers
        return conv_x, activated_x, pooled_x
    
# instantiate the model and set the weights
weight = torch.from_numpy(filters).unsqueeze(1).type(torch.FloatTensor)
model = Net(weight)

# print out the layer in the network
print(model)

Net(
  (conv): Conv2d(1, 4, kernel_size=(4, 4), stride=(1, 1), bias=False)
  (pool): MaxPool2d(kernel_size=4, stride=4, padding=0, dilation=1, ceil_mode=False)
)

The output of each filter is visualized

First, we will define a secondary function viz_layer, it receives a specific layer, and a plurality of filters (optional) as an input, an output and display of the layer after the image through.

# helper function for visualizing the output of a given layer
# default number of filters is 4
def viz_layer(layer, n_filters= 4):
    fig = plt.figure(figsize=(20, 20))
    
    for i in range(n_filters):
        ax = fig.add_subplot(1, n_filters, i+1, xticks=[], yticks=[])
        # grab layer outputs
        ax.imshow(np.squeeze(layer[0,i].data.numpy()), cmap='gray')
        ax.set_title('Output %s' % str(i+1))

Let us look at the application ReLu function is activated, the convolution output layer is.

# plot original image
plt.imshow(gray_img, cmap='gray')

# visualize all filters
fig = plt.figure(figsize=(12, 6))
fig.subplots_adjust(left=0, right=1.5, bottom=0.8, top=1, hspace=0.05, wspace=0.05)
for i in range(4):
    ax = fig.add_subplot(1, 4, i+1, xticks=[], yticks=[])
    ax.imshow(filters[i], cmap='gray')
    ax.set_title('Filter %s' % str(i+1))

    
# convert the image into an input Tensor
gray_img_tensor = torch.from_numpy(gray_img).unsqueeze(0).unsqueeze(1)

# get all the layers 
conv_layer, activated_layer, pooled_layer = model(gray_img_tensor)

# visualize the output of the activated conv layer
viz_layer(activated_layer)

The pooled visual output layer

Then, look at the output of the cell layer. Pooling layer depicted above feature map as an input, and through a number of cell factor, through a given core region configured to have only a maximum new smaller image (i.e., the brightest) to reduce the value of those map dimensions.

# visualize the output of the pooling layer
viz_layer(pooled_layer)