A Guide to Interpretability in Image Classification Techniques

Contents

• 1. Introduction
  • 1.2 About the book
• 2. What is Computer Vision?
  • 2.1 Image Classification
    • 2.1.1 What is Image Classification?
    • 2.1.2 Applications of Image Classification Algorithms
    • 2.1.3 Challenges faced by Image Classification Algorithms
  • 2.2 Object Detection
    • 2.2.1 What is Object Detection?
    • 2.2.2 Applications of Object Detection Algorithms
    • 2.2.3 Challenges faced by Object Detection Algorithms
  • 2.3 Image Segmentation
    • 2.3.1 What is Image Segmentation?
    • 2.3.2 Applications of Image Segmentation Algorithms
    • 2.3.3 Challenges faced by Image Segmentation Algorithms
  • 2.4 Pose Estimation
    • 2.4.1 What is Pose Estimation?
    • 2.4.2 Applications of Pose Estimation Algorithms
    • 2.4.3 Challenges faced by Pose Estimation Algorithms
  • 2.5 Image Captioning
    • 2.5.1 What is Image Captioning?
    • 2.5.2 Applications of Image Captioning Algorithms
    • 2.5.3 Challenges faced by Image Captioning Algorithms
  • 2.6 Image Synthesis
    • 2.6.1 What is Image Synthesis?
    • 2.6.2 Applications of Image Synthesis Algorithms
    • 2.6.3 Challenges of Image Synthesis Algorithms
• 3. Understanding Interpretability
  • 3.1 What is Interpretability?
  • 3.2 Understanding Explainability and its relation with Interpretability
    • 3.2.1 Important Properties Of Explainability
  • 3.3 Why Interpretability is necessary?
    • 3.3.1 Verify that the classifier works correctly
    • 3.3.2 Improve the classifier
    • 3.3.3 Learn from the algorithm itself about it’s decisions
    • 3.3.4 Get insights
• 4. Interpretability in Computer Vision
  • 4.1 Grad-CAM
  • 4.2 Grad-CAM++
• 5. Observations and Conclusion