search

weizhou-geek / VGCN-PyTorch

PyTorch Implementation of TCSVT 2020 "Blind Omnidirectional Image Quality Assessment with Viewport Oriented Graph Convolutional Networks"
MIT License
23 stars 8 forks source link

"Affinity Matrix" Question #2

Closed dothuhahb98 closed 3 years ago

dothuhahb98 commented 3 years ago

Thank you for you distribution.

I am trying to test your model in my dataset. In your paper, you mentioned about how to calculate the Affinity matrix which was represented in matrix (20x20). I am trying to calculate by using Matlab, but I have some misunderstanding about the input of it. Can you clarifies it and give some tutorial about the how you create the Affinity matrix.

Thanks a lot.

weizhou-geek commented 3 years ago

When we get the selected viewports, if the distance between center point of viewport A and viewport B is larger than 45°, then the edge is between the two nodes is 0, otherwise, the edge is set as 1. We can obtain the location of center point in fov_selection folder.

dothuhahb98 commented 3 years ago

I got it. Thanks a lot.