Convolutions

Depthwise Convolution

Depthwise Convolution is a type of convolution where we apply a single convolutional filter for each input channel. In the regular 2D convolution performed over multiple input channels, the filter is as deep as the input and lets us freely mix channels to generate each element in the output. In contrast, depthwise convolutions keep each channel separate. To summarize the steps, we:

  1. Split the input and filter into channels.
  2. We convolve each input with the respective filter.
  3. We stack the convolved outputs together.

Image Credit: Chi-Feng Wang

Papers


Paper Code Results Date Stars

Tasks


Task Papers Share
Image Classification 100 13.81%
Object Detection 68 9.39%
Semantic Segmentation 49 6.77%
General Classification 45 6.22%
Quantization 26 3.59%
Instance Segmentation 16 2.21%
Model Compression 16 2.21%
Knowledge Distillation 10 1.38%
Network Pruning 10 1.38%

Components


Component Type
🤖 No Components Found You can add them if they exist; e.g. Mask R-CNN uses RoIAlign

Categories