YOLACT++: Better Real-time Instance Segmentation

3 Dec 2019Daniel BolyaChong ZhouFanyi XiaoYong Jae Lee

We present a simple, fully-convolutional model for real-time (>30 fps) instance segmentation that achieves competitive results on MS COCO evaluated on a single Titan Xp, which is significantly faster than any previous state-of-the-art approach. Moreover, we obtain this result after training on only one GPU... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) mask AP 34.6% # 14
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) AP50 53.8 # 7
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) AP75 36.9 # 5
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) APS 11.9 # 8
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) APM 36.8 # 6
Instance Segmentation COCO test-dev YOLACT-550++ (ResNet-101-FPN) APL 55.1 # 4
Real-time Instance Segmentation MSCOCO YOLACT-550++ (ResNet-101-FPN) AP 34.6 # 3
Real-time Instance Segmentation MSCOCO YOLACT-550++ (ResNet-101-FPN) Frame (fps) 27.3 # 3