Limitation of Feed-Forward Neural Network and CNN :
Issues with Feed-Forward Neural Network:
- Loss of neighborhood information.
- More parameters to optimize.
- It’s not Translation invariance
Issues with Convolutional Neural Network:
- CNN is a translation invariance but it’s not translation equivariance.
- Compared to NN it having fewer parameters to train.
To overcome these limitations we use a capsule network i.e introduce by Geoffrey Hinton.