Semantic Segmentation Models
Fully Convolutional Network
Introduced by Shelhamer et al. in Fully Convolutional Networks for Semantic SegmentationFully Convolutional Networks, or FCNs, are an architecture used mainly for semantic segmentation. They employ solely locally connected layers, such as convolution, pooling and upsampling. Avoiding the use of dense layers means less parameters (making the networks faster to train). It also means an FCN can work for variable image sizes given all connections are local.
The network consists of a downsampling path, used to extract and interpret the context, and an upsampling path, which allows for localization.
FCNs also employ skip connections to recover the fine-grained spatial information lost in the downsampling path.
Source: Fully Convolutional Networks for Semantic Segmentation
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Semantic Segmentation | 94 | 18.25% |
| Image Segmentation | 41 | 7.96% |
| Object Detection | 23 | 4.47% |
| Medical Image Segmentation | 20 | 3.88% |
| General Classification | 20 | 3.88% |
| Image Classification | 15 | 2.91% |
| Classification | 9 | 1.75% |
| Time Series Classification | 7 | 1.36% |
| Denoising | 7 | 1.36% |
Usage Over Time
Components
| Component | Type |
|
|---|---|---|
Convolution
|
Convolutions | |
Dropout
|
Regularization | (optional) |
Max Pooling
|
Pooling Operations |
