UVO is a new benchmark for open-world class-agnostic object segmentation in videos.
23 PAPERS • 3 BENCHMARKS
Fetoscopic Placental Vessel Segmentation and Registration (FetReg2021) challenge was organized as part of the MICCAI2021 Endoscopic Vision (EndoVis) challenge. The dataset is useful for the development of generalized and robust semantic segmentation and video mosaicking algorithms for long duration fetoscopy videos.
2 PAPERS • NO BENCHMARKS YET
…We benchmark four foundational video understanding tasks: action recognition, action segmentation, object detection and multi-object tracking.
1 PAPER • NO BENCHMARKS YET
…Subtask 2: Stroke Classification: Participants are required to build a classification system that automatically labels video segments according to a performed stroke.
3 PAPERS • 2 BENCHMARKS