SERNet-Former: Semantic Segmentation by Efficient Residual Network with Attention-Boosting Gates and Attention-Fusion Networks

Benchmark Model Rank Results
semantic-segmentation-on-ade20kSERNet-Former#17Validation mIoU: 59.35
semantic-segmentation-on-ade20k-valSERNet-Former_v2#8mIoU: 59.35
semantic-segmentation-on-bdd100k-valSERNet-Former_v2#2mIoU: 67.42
semantic-segmentation-on-camvidSERNet-Former#1Mean IoU: 84.62
semantic-segmentation-on-cityscapesSERNet-Former#6Mean IoU (class): 84.83
semantic-segmentation-on-cityscapes-valSERNet-Former#2mIoU: 87.35Validation mIoU: 87.35