HyperAI초신경
홈뉴스최신 연구 논문튜토리얼데이터셋백과사전SOTALLM 모델GPU 랭킹컨퍼런스
전체 검색
소개
HyperAI초신경

Video Panoptic Segmentation On Vipseg

평가 지표

STQ
VPQ

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름
STQ
VPQ
Paper TitleRepository
TarViS (ResNet-50)43.133.5TarViS: A Unified Approach for Target-based Video Segmentation
CAVIS(VIT-L)56.158.5Context-Aware Video Instance Segmentation
UniVS(Swin-L)58.249.3UniVS: Unified and Universal Video Segmentation with Prompts as Queries
Tube-Link(Swin-base)49.450.4Tube-Link: A Flexible Cross Tube Framework for Universal Video Segmentation
DVIS(Swin-L)55.357.6DVIS: Decoupled Video Instance Segmentation Framework
DEVA (Mask2Former - SwinB)52.255.0Tracking Anything with Decoupled Video Segmentation
MaXTron (ConvNeXt-L)-57.1A Simple Video Segmenter by Tracking Objects Along Axial Trajectories
MaXTron (ConvNeXtV2-L)-58.0A Simple Video Segmenter by Tracking Objects Along Axial Trajectories
TarViS (Swin-L)52.948.0TarViS: A Unified Approach for Target-based Video Segmentation
DVIS++(VIT-L)56.058.0DVIS++: Improved Decoupled Framework for Universal Video Segmentation
MaXTron (ResNet-50)-46.7A Simple Video Segmenter by Tracking Objects Along Axial Trajectories
TarViS (Swin-T)45.335.8TarViS: A Unified Approach for Target-based Video Segmentation
0 of 12 row(s) selected.