Hi Caffe Users!!!
I was referring latest materials on deep learning for deploying them to real time systems......
Came across this e-pub
Has anyone tried these Deep Network Cascades ....for now I was planning to skip the Fast Feature Cascade aside....
Had few questions on following:
1. Few hints on selecting architecture of cascades
2. Fixing the kernel size and the depth of each layer
3. Any info on using output predictions from one network to filter input to next network....
Would like to discuss with anyone interested to implement/has implemented this....
Cheers!!!