(PAMI 2017) DeepLab:Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

Keyword [DeepLabv2] [Dilated Conv] [ASPP] [CRF]

Chen L C, Papandreou G, Kokkinos I, et al. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 40(4): 834-848.

1. Overview

In this paper, it proposes DeepLabv2 for segmentation.

  • Dilated Conv (solve challenge: reduced feature resolution).
  • Astrous Spatial Pyramid Pooling (solve challenge: multiple scales).
  • CRF (sovel challenge: reduced localization accuracy)

1.1. Dilated Conv

1.2. Astrous Spatial Pyramid Pooling (ASPP)

1.3. Details