SE-DETR: Explicit Semantic Exploration for Generalizability and Distinguishability in Video Temporal Grounding

Benchmark Model Rank Results
natural-language-moment-retrieval-on-tacosSE-DETR+#12R@1,IoU=0.3: 50.6R@1,IoU=0.5: 39.2R@1,IoU=0.7: 22.6mIoU: 36.2
moment-retrieval-on-qvhighlightsSE-DETR+#16mAP: 46.42R@1 IoU=0.5: 68.22R@1 IoU=0.7: 49.96mAP@0.5: 69.74…
moment-retrieval-on-charades-staSE-DETR+#31R@1 IoU=0.5: 60.2R@1 IoU=0.7: 39.7R@1 IoU=0.3: 71.1mIoU: 51.2