Global-Local Processing in Convolutional Neural Networks
Abstract: Convolutional Neural Networks (CNNs) have achieved outstanding performance on image processing challenges. Actually, CNNs imitate the typically developed human brain structures at the micro-level (Artificial neurons). At the same time, they distance themselves from imitating natural visual perception in humans at the macro architectures (high-level cognition). Recently it has been investigated that CNNs are highly biased toward local features and fail to detect the global aspects of their input. Nevertheless, the literature offers limited clues on this problem. To this end, we propose a simple yet effective solution inspired by the unconscious behavior of the human pupil. We devise a simple module called Global Advantage Stream (GAS) to learn and capture the holistic features of input samples (i.e., the global features). Then, the GAS features were combined with a CNN network as a plug-and-play component called the Global/Local Processing (GLP) model. The experimental results confirm that this stream improves the accuracy with an insignificant additional computational/temporal load and makes the network more robust to adversarial attacks. Furthermore, investigating the interpretation of the model shows that it learns a more holistic representation similar to the perceptual system of healthy humans
- Reveal of vision transformers robustness against adversarial attacks. arXiv preprint arXiv:2106.03734.
- Visual scanning and pupillary responses in young children with autism spectrum disorder. Journal of Clinical and Experimental Neuropsychology, 28(7):1238–1256.
- Deep convolutional networks do not classify based on global object shape. PLoS computational biology, 14(12):e1006613.
- Towards evaluating the robustness of neural networks. In 2017 ieee symposium on security and privacy (sp), pages 39–57. IEEE.
- Autism spectrum disorder and pupillometry: A systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews, 120:479–508.
- An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929.
- Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories. Computer Vision and Pattern Recognition Workshop.
- Deep learning. MIT press.
- Explaining and harnessing adversarial examples. arXiv preprint arXiv:1412.6572.
- Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770–778.
- The origins and prevalence of texture bias in convolutional neural networks. Advances in Neural Information Processing Systems, 33:19000–19015.
- Adversarial examples in the physical world. In Artificial intelligence safety and security, pages 99–112. Chapman and Hall/CRC.
- Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive psychology, 9(3):353–383.
- Technical report on the cleverhans v2. 1.0 adversarial examples library. arXiv preprint arXiv:1610.00768.
- Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32.
- Biologically inspired mechanisms for adversarial robustness. arXiv preprint arXiv:2006.16427.
- Task-induced pupil response and visual perception in adults. PloS one, 13(12):e0209556.
- Grad-cam: Visual explanations from deep networks via gradient-based localization. In Proceedings of the IEEE international conference on computer vision, pages 618–626.
- Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 1–9.
- Pupillometry reveals perceptual differences that are tightly linked to autistic traits in typical adults. Elife, 7:e32399.
- Attention is all you need. Advances in neural information processing systems, 30.
- Wandell, B. A. (1995). Foundations of vision. Sinauer Press,.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.
