Image-to-text R@1: 72.3Text-to-image R@1: 72.6
Papers with Code · CC BY-SA 4.0
Cross-Modal Retrieval
Cross-Modal Retrieval (CMR) is a task of retrieving items across different modalities, such as image, text, video, and audio. The core challenge of CMR is the heterogeneity gap, which arises because data from different modalities have distinct representations, making direct comparison difficult. To address this, most CMR methods focus on learning a shared latent embedding space. In this space, concepts from different modalities are projected, allowing their similarity to be measured using a distance metric.
<span class="description-source">Scene-centric vs. Object-centric Image-Text Cross-modal Retrieval: A Reproducibility Study</span>
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
X2-VLM (large)
Image-to-text R@1: 98.8Image-to-text R@10: 100Image-to-text R@5: 100Text-to-image R@1: 91.8
논문
코드
ReCon
Image-to-text R@1: 80.9Image-to-text R@5: 96.6Image-to-text R@10: 98.8Text-to-image R@1: 65.2
논문
코드
ReCon
Image-to-text R@1: 80.3Image-to-text R@5: 95.3Image-to-text R@10: 97.8Text-to-image R@1: 61.6
논문
코드
ReCon
Image-to-text R@1: 43.1Image-to-text R@5: 68.7Image-to-text R@10: 78.1Text-to-image R@1: 44.9
논문
코드
데이터셋
18개논문
Image-to-text R@1: 74.9Text-to-image R@1: 75.6
Image-to-text R@1: 54.0Text-to-image R@1: 54.9
Image-to-text R@1: 64Text-to-image R@1: 63.9
Image-to-text R@1: 39.8Text-to-image R@1: 40.2
Image-to-text R@1: 41.2Text-to-image R@1: 25.3
Image-to-text R@1: 51.8Text-to-image R@1: 52.8
Image-to-text R@1: 81.9Text-to-image R@1: 66.9
Image-to-text R@1: 60.0Text-to-image R@1: 60.3
Image-to-text R@1: 98.8Image-to-text R@10: 100Image-to-text R@5: 100Text-to-image R@1: 91.8
라이브러리
Papers with Code 리더보드 row에 연결된 구현 저장소입니다.