Spectrum-guided Multi-granularity Referring Video Object Segmentation

Benchmark Model Rank Results
referring-expression-segmentation-on-a2dSgMg (Video-Swin-B)#1AP: 0.585IoU overall: 0.799IoU mean: 0.720Precision@0.5: 0.843
referring-expression-segmentation-on-davisSgMg#5J&F 1st frame: 63.3
referring-expression-segmentation-on-j-hmdbSgMg (Video-Swin-B)#1AP: 0.450IoU overall: 0.737IoU mean: 0.725Precision@0.5: 0.972
referring-expression-segmentation-on-refer-1SgMg (Pre-training)#12J&F: 65.7J: 63.9F: 67.4
referring-video-object-segmentation-on-refSgMg#7J&F: 63.3J: 60.6F: 66.0
referring-video-object-segmentation-on-referSgMg#7J&F: 65.7J: 63.9F: 67.4