D2Conv3D: Dynamic Dilated Convolutions for Object Segmentation in Videos
Open paper
Benchmark
Model
Rank
Results
unsupervised-video-object-segmentation-on-davis-2016-val
D2Conv3D
#7
G: 86.0
J: 85.5
F: 86.5
video-instance-segmentation-on-ovis-validation
D2Conv3D (ResNet-50)
#38
mask AP: 15.2
AP50: 33.8
AP75: 13.7
Rank counts only results with a code link.