Generally, we come across terms such as High Frequency and Low frequency filters in Convolutional Neural Networks (CNN). In regards to this highlighted statement, in 'S1' section of this paper by Jason Yosinski (ref 1), enter image description here I thought, in order for high and low-frequency filters to produce a similar effect, weight of low-frequency filters should be greater than high-frequency filters. I would like to understand why I am wrong and I will be grateful if anyone can elaborate about High and Low frequency filters in CNN, in general, or in this context. Thank you.

Ref 1: Yosinski, Jason, et al. "Understanding neural networks through deep visualization." arXiv preprint arXiv:1506.06579 (2015).



You must log in to answer this question.

Browse other questions tagged .