Pointwise Convolution is a type of convolution that uses a 1x1 kernel: a kernel that iterates through every single point. This kernel has a depth of however many channels the input image has. It can be used in conjunction with depthwise convolutions to produce an efficient class of convolutions known as depthwise-separable convolutions.
Image Credit: Chi-Feng Wang
Paper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Image Classification | 70 | 10.54% |
Object Detection | 48 | 7.23% |
Classification | 37 | 5.57% |
Semantic Segmentation | 32 | 4.82% |
Quantization | 30 | 4.52% |
Instance Segmentation | 11 | 1.66% |
Management | 9 | 1.36% |
Ensemble Learning | 8 | 1.20% |
Computational Efficiency | 8 | 1.20% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |