PK PyTorchKR Papers
Papers with Code · CC BY-SA 4.0

Semi-Supervised Video Object Segmentation

Video Object Segmentation › Semi-Supervised Video Object Segmentation 382 평가 기록 18 datasets

The semi-supervised scenario assumes the user inputs a full mask of the object(s) of interest in the first frame of a video sequence. Methods have to produce the segmentation mask for that object(s) in the subsequent frames.

개요는 각 항목의 일부입니다. 위 탭에서 전체 기록을 페이지별로 탐색할 수 있습니다.

벤치마크

Papers with Code 아카이브의 데이터셋별 기록 예시입니다. 평가 조건과 지표 방향이 검증되지 않아 순위를 의미하지 않습니다.

Dataset기록된 모델 예시MetricPaperCode
SAM2
J&F: 90.7Params(M): 224.4
논문 코드
ISVOS (BL30K, MS)
Jaccard (Mean): 92.5F-measure (Mean): 94.2J&F: 93.4
논문 -
Cutie+ (base, MEGA)
J&F: 88.1Jaccard (Mean): 84.7F-measure (Mean): 91.4FPS: 17.9
논문 코드
Cutie+ (base, MEGA)
F-Measure (Seen): 91.0F-Measure (Unseen): 90.1Overall: 87.5Jaccard (Seen): 86.6
논문 코드
HMMN
FPS: 10.0D16 val (G): 89.4D16 val (J): 88.2D16 val (F): 90.6
논문 코드
Cutie+ (base, MEGA)
Overall: 87.5Jaccard (Seen): 86.3Jaccard (Unseen): 82.7F-Measure (Seen): 90.6
논문 코드
DAM4SAM
EAO: 0.729
논문 코드
SAM2
J&F: 77.9
논문 코드
ISVOS
J&F: 90.0J: 88.3F: 91.7
논문 -
Spatiotemporal CNN
mIoU: 0.796
논문 코드
Cutie (base, MEGA, 600 pixels)
HOTA (all): 66.0HOTA (common): 66.5HOTA (uncommon): 65.9
논문 코드
Cutie (base, MEGA, 600 pixels)
HOTA (all): 61.2HOTA (common): 65.0HOTA (uncommon): 60.3
논문 코드

데이터셋

18개
DAVIS 2017 (val)DAVIS 2016DAVIS 2017 (test-dev)YouTube-VOS 2018DAVIS (no YouTube-VOS training)YouTube-VOS 2019VOT2020MOSELong Video DatasetYouTubeBURST-testBURST-val

논문

라이브러리

Papers with Code 리더보드 row에 연결된 구현 저장소입니다.

github.com/hkchengrex/Cutie 1 논문
github.com/hkchengrex/Tracking-Anything-with-DEVA 1 논문
github.com/bowang-lab/medsam2 1 논문
github.com/dcnieho/segment-anything-2 1 논문
github.com/facebookresearch/sam2 1 논문
github.com/facebookresearch/segment-anything 1 논문
github.com/ibaiGorordo/ONNX-SAM2-Segment-Anything 1 논문
github.com/idea-research/grounded-sam-2 1 논문

하위 태스크

OV One-shot visual object segmentation 2 데이터셋 · 3 rows