Proceedings

Multi-view Feature Augmentation with Adaptive Class Activation Mapping

Xiang Gao, Yingjie Tian, Zhiquan Qi

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 678-684.

https://doi.org/10.24963/ijcai.2021/94

Keywords

  • Computer Vision: Recognition: Detection, Categorization, Indexing, Matching, Retrieval, Semantic Interpretation
  • Data Mining: Classification

Abstract

We propose an end-to-end-trainable feature augmentation module built for image classification that extracts and exploits multi-view local features to boost model performance. Different from using global average pooling (GAP) to extract vectorized features from only the global view, we propose to sample and ensemble diverse multi-view local features to improve model robustness. To sample class-representative local features, we incorporate a simple auxiliary classifier head (comprising only one 1x1 convolutional layer) which efficiently and adaptively attends to class-discriminative local regions of feature maps via our proposed AdaCAM (Adaptive Class Activation Mapping). Extensive experiments demonstrate consistent and noticeable performance gains achieved by our multi-view feature augmentation module.

Xiang Gao, Yingjie Tian, and Zhiquan Qi. 2021. Multi-view Feature Augmentation with Adaptive Class Activation Mapping. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 678–684. https://doi.org/10.24963/ijcai.2021/94
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards