Improving Semantic Segmentation via Video Propagation and Label Relaxation

CVPR 2019 Yi ZhuKaran SapraFitsum A. RedaKevin J. ShihShawn NewsamAndrew TaoBryan Catanzaro

Semantic segmentation requires large amounts of pixel-wise annotations to learn accurate models. In this paper, we present a video prediction-based methodology to scale up training sets by synthesizing new training samples in order to improve the accuracy of semantic segmentation networks... (read more)

PDF Abstract

Evaluation Results from the Paper


 SOTA for Semantic Segmentation on CamVid (using extra training data)

     Get a GitHub badge
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK USES EXTRA
TRAINING DATA
COMPARE
Semantic Segmentation CamVid DeepLabV3Plus + SDCNetAug Mean IoU 81.7% # 1
Semantic Segmentation Cityscapes test DeepLabV3Plus + SDCNetAug Mean IoU (class) 83.5% # 4
Semantic Segmentation KITTI Semantic Segmentation DeepLabV3Plus + SDCNetAug Mean IoU (class) 72.8% # 1