News for hackers. Stuff that matters.
Re:ImageNet-trained CNNs are biased towards texture (2019) by mbeex on Monday April 20, 2020 @04:16AM (#22922111)
I think, this is one of the reasons to use much more video-based training sets and these in time-correct frame order. A moving boundary is something quite significant, but these should possibly not exclusively inspected through a hole ('convolutinal')
Re:ImageNet-trained CNNs are biased towards texture (2019) by rayuela on Monday April 20, 2020 @10:58AM (#22924363)
Oh wow you're absolutely right. This is actually a really clever solution to the problem presented in this paper.
Re:ImageNet-trained CNNs are biased towards texture (2019) by mbeex on Monday April 20, 2020 @04:21AM (#22922118)
Oh, forgot an 'o'.
Re:ImageNet-trained CNNs are biased towards texture (2019) by mkl on Monday April 20, 2020 @05:48AM (#22922477)
Did you know you can edit your comments for two hours? Click "edit" above the comment.

https://github.com/minimaxir/hacker-news-undocumented

Re:ImageNet-trained CNNs are biased towards texture (2019) by mbeex on Monday April 20, 2020 @11:30AM (#22924741)
This is actually new to me. Thank you!