SynMultiVid: A Synthetic Multi-Modal Video Dataset for Cross-Modal Correspondence Evaluation
收藏IEEE2026-04-17 收录
下载链接：
https://ieee-dataport.org/documents/synmultivid-synthetic-multi-modal-video-dataset-cross-modal-correspondence-evaluation-0
下载链接
链接失效反馈官方服务：
资源简介：
This dataset provides synchronized multi-modal videos with ground-truth correspondences for evaluating cross-modal video matching and registration methods. Starting from standard RGB videos, four complementary pseudo-modalities are generated to emulate typical cross-modal discrepancies: Motion Thermal (motion-sensitive responses), Temporal Gradient (time-based intensity variations), Frequency Domain Filtering (edge-emphasized structures), and Texture Suppression (geometry-preserving, texture-reduced representations).
提供机构：
Jie Wang; Caoqi Wei; CHEN YE GAN
