将CNN的输入由RGB 3通道变成192个频域通道，结果会如何？阿里达摩院新出的paper《Learning in the Frequency Domain》实践了这个idea，并且在图像分类、实例分割任务上取得不错的表现。

Specifically for ImageNet clas- sification with the same input size, the proposed method achieves 1.41% and 0.66% top-1 accuracy improvements on ResNet-50 and MobileNetV2, respectively. Even with half input size, the proposed method still improves the top-1 accuracy on ResNet-50 by 1%. In addition, we observe a 0.8% average precision improvement on Mask R-CNN for instance segmentation on the COCO dataset.

要点：

• 优势：降低输入的size，减少传统CNN第一层下采样丢失的信息

• 算法流程如下：
  RGB图->YCbCr->DCT-> 频域通道重排->频域通道选择->CNN

  
  DCT.png

• 用选择的频域通道直接替换layer0的下采样操作：

  
  resnet.png

• 通道选择通过一个小网络实现，直接把loss加到acc loss上

  
  select.png

原文及解读

• paper
• 解读