search

LiWentomng / Point2Mask

The code for "Point2Mask: Point-supervised Panoptic Segmentation via Optimal Transport", ICCV2023
Apache License 2.0
63 stars 6 forks source link

How does the model supervise the process of Semantic Decoder and Boundary Decoder? #5

Open jahoyan opened 7 months ago

jahoyan commented 7 months ago

Hello, I have a question about your model, that is, how does the model supervise the process of Semantic Decoder and Boundary Decoder?How are parameters updated in the backbone,Semantic Decoder and Boundary Decoder?

jahoyan commented 7 months ago

Hello, I have another question that you used OTP to match the gt_point and the sampling point.But in your config code you use HungarianAssigner to match the gt_point and the sampling point.So I want to know what's the difference between the two methods

LiWentomng commented 7 months ago

Hello, I have another question that you used OTP to match the gt_point and the sampling point.But in your config code you use HungarianAssigner to match the gt_point and the sampling point.So I want to know what's the difference between the two methods

@jahoyan
OT is used for generating pseudo mask. Based on the pseudo mask, the HungarianAssigner is used to assign pos/neg samples same to original fully supervised method.

LiWentomng commented 7 months ago

Hello, I have a question about your model, that is, how does the model supervise the process of Semantic Decoder and Boundary Decoder?How are parameters updated in the backbone,Semantic Decoder and Boundary Decoder?

@jahoyan
The parameters in Semantic Decoder and Boundary Decoder are driven by the weakly supervised loss, inlcuding semantic loss and boundary loss. Please see the Sec.3.4.1 in the paper.