PointDecoder

[BlouseDong]

Hello, I noticed that in your paper, it is mentioned that the point decoder is based on HRNet. However, looking at the code, the structure of PointDecoder combines a Transformer, embedding layers, and convolutional layers to process input features and generate point predictions. This structure is not based on HRNet.

[Hzzone]

Please see sec 3.3: the point decoder shares the same architecture as the mask decoder... You may misunderstand architecture and loss function.