/CNN-Visualization

TensorFlow implementations of visualization of convolutional neural networks, such as Grad-Class Activation Mapping and guided back propagation

Primary LanguagePythonMIT LicenseMIT

Visualization of Deep Covolutional Neural Networks

  • This repository contains implementations of visualizatin of CNN in recent papers.
  • The source code in the repository can be used to demostrate the algorithms as well as test on your own data.

Requirements

Algorithms

Visulization of filters and feature maps of GoogLeNet

  • The most straightforward approach to visualize a CNN is to show the feature maps (activations) and filters.
  • Details of the implementation and more results can be found here

Deconvnet

  • Pick a specific activation on a feature map and set other activation to zeros, then reconstruct an image by mapping back this new feature map to input pixel space.
  • Details of the implementation and more results can be found here. Some results:

Guided back propagation

  • Details of the implementation and more results can be found here. Some results:

gbp

Class Activation Mapping (CAM)

  • The class activation map highlights the most informative image regions relevant to the predicted class. This map can be obtained by adding a global average pooling layer at the end of convolutional layers.
  • Details of the implementation and more results can be found here. Some results:

celtech_change

Gradient-weighted Class Activation Mapping (Grad-CAM)

  • Grad-CAM generates similar class heatmap as CAM, but it does not require to re-train the model for visualizatin.
  • Details of the implementation and more results can be found here. Some results:

grad-cam-result