ResearchBenchmarksQuery-based video segmentation (Audio→Text) on AVSBench S4 VGGSound-AVEL 40KFollow74.3mIoUCoDAAR73.2673.5373.874.07May 12, 2026Evaluation ResultsMethodMethodLinksmIoUF-scoreCoDAARSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0574.384.5MICUSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0573.884.1DCIDSetting=AVT, Fine-tune...Setting=AVT, Fine-tuned epoch=4th2026.0573.383.3