HarDNet-MSEG: A Simple Encoder-Decoder Polyp Segmentation Neural Network that Achieves over 0.9 Mean Dice and 86 FPS

18 Jan 2021  ·  Chien-Hsiang Huang, Hung-Yu Wu, Youn-Long Lin ·

We propose a new convolution neural network called HarDNet-MSEG for polyp segmentation. It achieves SOTA in both accuracy and inference speed on five popular datasets... For Kvasir-SEG, HarDNet-MSEG delivers 0.904 mean Dice running at 86.7 FPS on a GeForce RTX 2080 Ti GPU. It consists of a backbone and a decoder. The backbone is a low memory traffic CNN called HarDNet68, which has been successfully applied to various CV tasks including image classification, object detection, multi-object tracking and semantic segmentation, etc. The decoder part is inspired by the Cascaded Partial Decoder, known for fast and accurate salient object detection. We have evaluated HarDNet-MSEG using those five popular datasets. The code and all experiment details are available at Github. https://github.com/james128333/HarDNet-MSEG read more

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Medical Image Segmentation CVC-ClinicDB HarDNet-MSEG mean Dice 0.9320 # 3
Medical Image Segmentation CVC-ColonDB HarDNet-MSED mean Dice 0.731 # 5
mIoU 0.660 # 5
Medical Image Segmentation ETIS-LARIBPOLYPDB HarDNet-MSED mIoU 0.613 # 5
mean Dice 0.677 # 5
Medical Image Segmentation Kvasir-SEG HarDNet-MSEG Average MAE 0.025 # 1
mean Dice 0.912 # 3
S-Measure 0.923 # 1
max E-Measure 0.958 # 1
mIoU 0.857 # 4
FPS 116 # 2

Methods