JM-CLIP: A Joint Modal Similarity Contrastive Learning Model for Video-Text Retrieval | IEEE Conference Publication | IEEE Xplore