BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation

ECCV 2018 Changqian YuJingbo WangChao PengChangxin GaoGang YuNong Sang

Semantic segmentation requires both rich spatial information and sizeable receptive field. However, modern approaches usually compromise spatial resolution to achieve real-time inference speed, which leads to poor performance... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Real-Time Semantic Segmentation CamVid BiSeNet mIoU 68.7% # 2
Semantic Segmentation CamVid BiSeNet Mean IoU 68.7% # 3
Semantic Segmentation Cityscapes test BiSeNet (ResNet-101) Mean IoU (class) 78.9% # 26
Real-Time Semantic Segmentation Cityscapes test BiSeNet mIoU 74.7% # 4
Real-Time Semantic Segmentation Cityscapes test BiSeNet Frame (fps) 65.5 # 4