gaudetcj

10 ms·

1. 1.▲
   by gaudetcj 8y ago

   Yes that is exactly what we do.
2. 2.▲
   by gaudetcj 8y ago

   One of the authors here. One thing about quaternion convolution is that you can write a color image into quaternion space by considering each channel as an imaginary axis. This lets the convolution act on the entire color space in a differe
3. 3.▲
   by gaudetcj 8y ago

   This paper was mostly to lay out the framework and give out the Keras layers for others to use. We expect that the biggest improvements will come from segmentation where the gains may come from treating each color channel as an imaginary ax
4. 4.▲
   by gaudetcj 8y ago

   Hi I'm one of the authors of this paper. Sorry if it unclear, but we reduce the number of filters per layer to account for this. The quaternion networks actually have fewer parameters.