Ablation-CAM: Visual Explanations for Deep Convolutional Network via Gradient-free Localization

By Saurabh Satish Desai, Harish Guruprasad Ramaswamy. IEEE Workshop/Winter Conference on Applications of Computer Vision, 2020

528 citations recorded by the source corpus.

In response to recent criticism of gradient-based visualization techniques, we propose a new methodology to generate visual explanations for deep Convolutional Neural Networks (CNN) - based models. Our approach – Ablation-based Class Activation Mapping (Ablation CAM) uses ablation analysis to determine the importance (weights) of individual feature map units w.r.t. class. Further, this is used to produce a coarse localization map highlighting the important regions in the image for predicting the concept. Our objective and subjective evaluations show that this gradient-free approach works better than state-of-the-art Grad-CAM technique. Moreover, further experiments are carried out to show that Ablation-CAM is class discriminative as well as can be used to evaluate trust in a model.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/9f5c43e85d2ae218f6082505dea8a5e980ed2f02

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Dejing Dou
2. Ruigang Fu
3. Haoyi Xiong
4. Lianghua He
5. Xiao Zhang
6. Xuanyu Wu
7. Xiaohu Dong
8. Jiang Bian

In Computer Vision and Pattern Recognition

1. Dejing Dou
2. Ruigang Fu
3. Haoyi Xiong
4. Lianghua He
5. Xiao Zhang
6. Xuanyu Wu
7. Xiaohu Dong
8. Jiang Bian

Citations in the corpus

• Interpretable deep learning: interpretation, interpretability, trustworthiness, and beyond (2021, background): “More works based on CAM can be found [24, 95, 110, 130, 142].”
• Gold-YOLO: Efficient Object Detector via Gather-and-Distribute Mechanism (2023, unknown): “(b) and (c) is AblationCAM [25] visualization”
• Axiom-based Grad-CAM: Towards Accurate Visualization and Explanation of CNNs (2020, methodology): “As an important branch of activation-based methods, CAM methods [3, 4, 17, 23] visualize CNN’s decision using feature maps of deep layers.”
• Decoupling Deep Learning for Interpretable Image Recognition (2022, methodology): “Ablation-CAM [21] analyzes the contribution of each factor to the network.”