F

FA_Label_VOS

Encoder decoder architecture for semantic segmentation .Encoder consists of a pretrained resnet18 and the decoder is composed of multiple layers of transposed convolution.