Question about Multi-view and Single-view Feature Interaction in Code (Figure 2) #21
Description
Thank you for providing the source code and sharing your excellent work. I have a few questions regarding the implementation.
In Figure 2 of your paper, it appears that
𝐶_𝑚 (Multi-view volume) and 𝐷_𝑠 (Single-view depth feature) are used to generate 𝐹_fuse.
However, upon examining the code, it seems that in the variable feature_mono_residual, Multi-view information is also included as part of the input feature alongside Single-view information.
Could you please clarify whether this is intended behavior? If so, could you explain how this aligns with the description in the paper?
Thank you for your time, and I look forward to your response.
