Image Model Blocks
Fractal Block
Introduced by Larsson et al. in FractalNet: Ultra-Deep Neural Networks without ResidualsA Fractal Block is an image model block that utilizes an expansion rule that yields a structural layout of truncated fractals. For the base case where $f_{1}\left(z\right) = \text{conv}\left(z\right)$ is a convolutional layer, we then have recursive fractals of the form:
$$ f_{C+1}\left(z\right) = \left[\left(f_{C}\circ{f_{C}}\right)\left(z\right)\right] \oplus \left[\text{conv}\left(z\right)\right]$$
Where $C$ is the number of columns. For the join layer (green in Figure), we use the element-wise mean rather than concatenation or addition.
Source: FractalNet: Ultra-Deep Neural Networks without Residuals
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Object Recognition | 1 | 33.33% |
| Translation | 1 | 33.33% |
| Image Classification | 1 | 33.33% |
Convolution
