search

kaix90 / DCTNet

460 stars 98 forks source link

Some differences between the code and paper #5

Open WXIAO-TJ opened 4 years ago

WXIAO-TJ commented 4 years ago

I'm confused about some differences between your code and paper. In the section 4.2 of your paper, you declare that you "pick the top 24 high-probability channels". I think the high-probability channels should follow the heat map in Fig 5. However, you pick the [0, 1, 2, 3, 8, 9, 10, 11, 16, 17, 18, 19, 24, 25, 26, 27] -th channels in your code which forms a square but doesn't obey the above distribution. I would appreciate it if you could help me out.

kaix90 commented 4 years ago

We tested the performance of several static patterns and we found the differences between these patterns are minor. In the code, we use the square pattern and it works pretty good. But you can also try other patterns. Thanks.

WXIAO-TJ commented 4 years ago

Thanks for your reply. But i still want to ask two questions:

  1. About above question, do you mean that the square pattern is better than the other patterns including the "Frequency Channel Selection"? 2.What is the order in which the 64 frequency channels are selected in Y, Cb, Cr. You just describe the order with the heat maps which is not comprehensive.