About the spatial_head in openpsg/models/roi_extractors/visiual_spatial.py

[shuizaola]

I am very interested in your work.I would really appreciate it if you could help me answer this question.
in line 516： roi_feats_result.append(head((roi_feats + rect_feats).view( roi_feats.size(0), -1))) ，I know that roi_feats represent the features of the union of subject and object, so what does rect_feats represent?
in line 474 :rect_input = torch.stack((head_rect, tail_rect),dim=1)
in line 477: rect_feats = self.spatial_conv(rect_input) What features does the rect_feats represent?（Does it represent the spatial featuresof subject and predicate? So what does this spatial feature mean?）

[Jingkang50]

Thanks for asking. If I remember it correctly, the spatial feature encodes the object location info by marking the localization of objects in interests in the original image with a rectangular bbox coordinate (or mask) and generate feature accordingly. The code is from MMSceneGraph, so we just use it. You might need to print and check rect_input to explore / learn on your own.

[shuizaola]

Thanks for asking. If I remember it correctly, the spatial feature encodes the object location info by marking the localization of objects in interests in the original image with a rectangular bbox coordinate (or mask) and generate feature accordingly. The code is from MMSceneGraph, so we just use it. You might need to print and check to explore / learn on your own.rect_input

Thanks you！