HyperAI초신경

홈 뉴스 최신 연구 논문 튜토리얼 데이터셋 백과사전 SOTA LLM 모델 GPU 랭킹 컨퍼런스

한국어

HyperAI초신경

Cross Modal Retrieval With Noisy 3

평가 지표

Image-to-text R@1

Image-to-text R@10

Image-to-text R@5

R-Sum

Text-to-image R@1

Text-to-image R@10

Text-to-image R@5

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Image-to-text R@1	Image-to-text R@10	Image-to-text R@5	R-Sum	Text-to-image R@1	Text-to-image R@10	Text-to-image R@5	Paper Title	Repository
NAC	80.3	98.5	96.2	524.5	63.2	96.0	90.3	NAC: Mitigating Noisy Correspondence in Cross-Modal Matching Via Neighbor Auxiliary Corrector	-
LNC	78.2	98.5	95.8	519.9	62.6	95.4	89.4	Learning with Noisy Correspondence	-
REPAIR	78.3	98.3	96.8	521.2	62.5	95.5	89.8	REPAIR: Rank Correlation and Noisy Pair Half-replacing with Memory for Noisy Correspondence	-
BiCro*	78.8	98.6	96.1	523.2	63.7	95.7	90.3	BiCro: Noisy Correspondence Rectification for Multi-modality Data via Bi-directional Cross-modal Similarity Consistency	-
SREM	78.5	98.8	96.8	524.1	63.8	95.8	90.4	Noisy Correspondence Learning with Self-Reinforcing Errors Mitigation	-
CRCL	79.6	98.7	96.1	525.6	64.7	95.9	90.6	Cross-modal Active Complementary Learning with Self-refining Correspondence
ReCon	80.9	98.8	96.6	528.6	65.2	96.0	91.0	ReCon: Enhancing True Correspondence Discrimination through Relation Consistency for Robust Noisy Correspondence Learning
L2RM-SCARF	80.2	98.5	96.3	524.7	64.2	95.4	90.1	Learning to Rematch Mismatched Pairs for Robust Cross-Modal Retrieval
UGNCL	79.5	99.0	97.2	526.3	63.7	96.0	90.9	UGNCL: Uncertainty-Guided Noisy Correspondence Learning for Efficient Cross-Modal Matching
CTPR-SGR	79.8	98.9	96.6	527	63.8	96.7	91.2	Learning From Noisy Correspondence With Tri-Partition for Cross-Modal Matching	-
DECL-SGARF	77.5	98.4	95.9	518.2	61.7	95.4	89.3	Deep Evidential Learning with Noisy Correspondence for Cross-Modal Retrieval
MSCN	78.1	98.8	97.2	524.6	64.3	95.8	90.4	Noisy Correspondence Learning with Meta Similarity Correction
LG-ITM-SGARF	79.6	98.5	96.5	524.9	64.4	95.9	90.0	Integrating Language Guidance Into Image-Text Matching for Correcting False Negatives
CREAM	78.9	98.6	96.3	523	63.3	95.8	90.1	Cross-modal Retrieval with Noisy Correspondence via Consistency Refining and Mining
RCL-SGR	77.0	98.1	95.5	515.5	61.3	94.8	88.8	Cross-Modal Retrieval with Partially Mismatched Pairs
GSC-SGR	79.5	98.9	96.4	525.7	64.4	95.9	90.6	Mitigating Noisy Correspondence by Geometrical Structure Consistency Learning
NCR	77.7	98.2	95.5	518.5	62.5	95.3	89.3	Learning with Noisy Correspondence for Cross-modal Matching

0 of 17 row(s) selected.