CA-Stream: Attention-based pooling for interpretable image recognition
Not on the map. Why?
Abstract (arXiv)
Explanations obtained from transformer-based architectures in the form of raw attention, can be seen as a class-agnostic saliency map. Additionally, attention-based pooling serves as a form of masking the in feature space. Motivated by this observation, we design an attention-based pooling mechanism intended to replace Global Average Pooling (GAP) at inference. This mechanism, called Cross-Attention Stream (CA-Stream), comprises a stream of cross attention blocks interacting with features at different network depths. CA-Stream enhances interpretability in models, while preserving recognition performance.
Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 49 of its 59 references that are in the map. First posted on arXiv 23 Apr 2024.
Builds on
49 of 59 references.
References · 49
- 2017Attention Is All You Needcited by 60,361
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
- 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
- 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
- 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,070
- 2022A ConvNet for the 2020scited by 3,579
- 2017Axiomatic Attribution for Deep Networkscited by 3,384
- 2015Learning Deep Features for Discriminative Localizationcited by 2,436
- 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,371
- 2017SmoothGrad: removing noise by adding noisecited by 1,109
- 2014Striving for Simplicity: The All Convolutional Netcited by 990
37 more
- 2018A Survey Of Methods For Explaining Black Box Modelscited by 973
- 2017Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networkscited by 939
- 2018RISE: Randomized Input Sampling for Explanation of Black-box Modelscited by 838
- 2017Network Dissection: Quantifying Interpretability of Deep Visual Representationscited by 708
- 2020Transformer Interpretability Beyond Attention Visualizationcited by 627
- 2017Interpretable Explanations of Black Boxes by Meaningful Perturbationcited by 612
- 2019Stand-Alone Self-Attention in Vision Modelscited by 542
- 2021LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inferencecited by 475
- 2021Early Convolutions Help Transformers See Bettercited by 396
- 2019Attention Augmented Convolutional Networkscited by 363
- 2021Rethinking Spatial Dimensions of Vision Transformerscited by 361
- 2017Right for the Right Reasons: Training Differentiable Models by Constraining their Explanationscited by 294
- 2020A Survey on Neural Network Interpretabilitycited by 291
- 2019Understanding Deep Networks via Extremal Perturbations and Smooth Maskscited by 259
- 2017Real Time Image Saliency for Black Box Classifierscited by 246
- 2014Object Detectors Emerge in Deep Scene CNNscited by 179
- 2021Conformer: Local Features Coupling Global Representations for Visual Recognitioncited by 164
- 2020Axiom-based Grad-CAM: Towards Accurate Visualization and Explanation of CNNscited by 151
- 2018Explaining Image Classifiers by Counterfactual Generationcited by 144
- 2020Restricting the Flow: Information Bottlenecks for Attributioncited by 110
- 2017Interpreting Deep Visual Representations via Network Dissectioncited by 103
- 2022A Consistent and Efficient Evaluation Strategy for Attribution Methodscited by 102
- 2021The Out-of-Distribution Problem in Explainability and Search Methods for Feature Importance Explanationscited by 87
- 2017Beyond Sparsity: Tree Regularization of Deep Models for Interpretabilitycited by 68
- 2021Improving Deep Learning Interpretability by Saliency Guided Trainingcited by 68
- 2018Local Explanation Methods for Deep Neural Networks Lack Sensitivity to Parameter Valuescited by 40
- 2022Metrics for saliency map evaluation of deep learning explanation methodscited by 40
- 2020SCOUTER: Slot Attention-based Classifier for Explainable Image Recognitioncited by 29
- 2021Augmenting Convolutional networks with attention-based aggregationcited by 29
- 2023Opti-CAM: Optimizing saliency maps for interpretabilitycited by 25
- 2019Saliency Learning: Teaching the Model Where to Pay Attentioncited by 23
- 2020Global Self-Attention Networks for Image Recognitioncited by 16
- 2019Regional Tree Regularization for Interpretability in Black Box Modelscited by 14
- 2021LFI-CAM: Learning Feature Importance for Better Visual Explanationcited by 13
- 2021Resisting Out-of-Distribution Data Problem in Perturbation of XAIcited by 12
- 2018Classifier-agnostic saliency map extractioncited by 7
- 2020Investigating and Simplifying Masking-based Saliency Methods for Model Interpretabilitycited by 7
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.