SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization
arXiv:2006.14255
Abstract
Interpretation of the underlying mechanisms of Deep Convolutional Neural Networks has become an important aspect of research in the field of deep learning due to their applications in high-risk environments. To explain these black-box architectures there have been many methods applied so the internal decisions can be analyzed and understood. In this paper, built on the top of Score-CAM, we introduce an enhanced visual explanation in terms of visual sharpness called SS-CAM, which produces centralized localization of object features within an image through a smooth operation. We evaluate our method on the ILSVRC 2012 Validation dataset, which outperforms Score-CAM on both faithfulness and localization tasks.
arXiv comment7 pages, 4 figures and 4 tables
References in corpus (14)
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- A Unified Approach to Interpreting Model Predictions
- SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
- Deep Residual Learning for Image Recognition
- Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networks
- Learning Important Features Through Propagating Activation Differences
- SmoothGrad: removing noise by adding noise
- Advancements in Image Classification using Convolutional Neural Network
- Learning Deep Features for Discriminative Localization
- Understanding Integrated Gradients with SmoothTaylor for Deep Neural Network Attribution
- Smoothed Geometry for Robust Attribution
- Hybrid coarse-fine classification for head pose estimation
- XDeep: An Interpretation Tool for Deep Neural Networks
Cited by in corpus (5)
- Deep Weakly-Supervised Learning Methods for Classification and Localization in Histology Images: A Survey
- IS-CAM: Integrated Score-CAM for axiomatic-based explanations
- SSA: Semantic Structure Aware Inference for Weakly Pixel-Wise Dense Predictions without Cost
- DiPS: Discriminative Pseudo-Label Sampling with Self-Supervised Transformers for Weakly Supervised Object Localization
- Efficient Modelling Across Time of Human Actions and Interactions