PraNet: Parallel Reverse Attention Network for Polyp Segmentation

13 Jun 2020  ·  Deng-Ping Fan, Ge-Peng Ji, Tao Zhou, Geng Chen, Huazhu Fu, Jianbing Shen, Ling Shao ·

Colonoscopy is an effective technique for detecting colorectal polyps, which are highly related to colorectal cancer. In clinical practice, segmenting polyps from colonoscopy images is of great importance since it provides valuable information for diagnosis and surgery. However, accurate polyp segmentation is a challenging task, for two major reasons: (i) the same type of polyps has a diversity of size, color and texture; and (ii) the boundary between a polyp and its surrounding mucosa is not sharp. To address these challenges, we propose a parallel reverse attention network (PraNet) for accurate polyp segmentation in colonoscopy images. Specifically, we first aggregate the features in high-level layers using a parallel partial decoder (PPD). Based on the combined feature, we then generate a global map as the initial guidance area for the following components. In addition, we mine the boundary cues using a reverse attention (RA) module, which is able to establish the relationship between areas and boundary cues. Thanks to the recurrent cooperation mechanism between areas and boundaries, our PraNet is capable of calibrating any misaligned predictions, improving the segmentation accuracy. Quantitative and qualitative evaluations on five challenging datasets across six metrics show that our PraNet improves the segmentation accuracy significantly, and presents a number of advantages in terms of generalizability, and real-time segmentation efficiency.

PDF Abstract

Results from the Paper


Ranked #4 on Camouflaged Object Segmentation on CAMO (using extra training data)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
Camouflaged Object Segmentation CAMO PraNet MAE 0.094 # 3
Weighted F-Measure 66.3 # 4
S-Measure 76.9 # 4
E-Measure 82.4 # 4
Medical Image Segmentation CVC-ClinicDB PraNet mean Dice 0.8990 # 23
Medical Image Segmentation CVC-ColonDB PraNet mean Dice 0.709 # 14
mIoU 0.649 # 13
Average MAE 0.045 # 1
S-Measure 0.819 # 5
max E-Measure 0.869 # 5
Medical Image Segmentation ETIS-LARIBPOLYPDB PraNet mIoU 0.5670 # 13
Average MAE 0.031 # 5
mean Dice 0.6280 # 13
S-Measure 0.794 # 4
max E-Measure 0.841 # 4
Medical Image Segmentation Kvasir-SEG PraNet Average MAE 0.030 # 8
mean Dice 0.898 # 21
S-Measure 0.915 # 6
max E-Measure 0.948 # 7
mIoU 0.849 # 19
Video Polyp Segmentation SUN-SEG-Easy (Unseen) PraNet S measure 0.733 # 6
mean E-measure 0.753 # 4
weighted F-measure 0.572 # 6
mean F-measure 0.632 # 7
Dice 0.621 # 8
Sensitivity 0.524 # 6
Video Polyp Segmentation SUN-SEG-Hard (Unseen) PraNet S-Measure 0.717 # 7
mean E-measure 0.735 # 7
weighted F-measure 0.544 # 6
mean F-measure 0.607 # 7
Dice 0.598 # 8
Sensitivity 0.512 # 6

Methods


No methods listed for this paper. Add relevant methods here