Loading...

Class attention map distillation for efficient semantic segmentation

Karimi Bavandpour, N ; Sharif University of Technology | 2020

369 Viewed
  1. Type of Document: Article
  2. DOI: 10.1109/MVIP49855.2020.9116875
  3. Publisher: IEEE Computer Society , 2020
  4. Abstract:
  5. In this paper, a novel method for capturing the information of a powerful and trained deep convolutional neural network and distilling it into a training smaller network is proposed. This is the first time that a saliency map method is employed to extract useful knowledge from a convolutional neural network for distillation. This method, despite of many others which work on final layers, can successfully extract suitable information for distillation from intermediate layers of a network by making class specific attention maps and then forcing the student network to mimic producing those attentions. This novel knowledge distillation training is implemented using state-of-the-art DeepLab and PSPNet segmentation networks and its effectiveness is shown by experiments on the standard Pascal Voc 2012 dataset. © 2020 IEEE
  6. Keywords:
  7. Knowledge Distillation ; Saliency Maps ; Computer vision ; Convolution ; Deep neural networks ; Distillation ; Distilleries ; Image segmentation ; Network layers ; Semantics ; Intermediate layers ; Saliency map ; Semantic segmentation ; State of the art ; Student network ; Convolutional neural networks
  8. Source: 1st International Conference on Machine Vision and Image Processing, MVIP 2020, 19 February 2020 through 20 February 2020 ; Volume 2020-February , 2020
  9. URL: https://ieeexplore.ieee.org/document/9116875