Finding Features with Deep learning auto-encoders.

Hi there, i have been training my neural network with autoencoders and it has been working well.

I followed this tutorial: https://www.mathworks.com/help/nnet/examples/training-a-deep-neural-network-for-digit-classification.html

I've got good results, 85% accuracy.

Now I want to isolate the region responsible for these predictions. I need to know which pixels in the image are contributing these predictions.

How do i go about doing that?

Thanks.

Answer 1

[Aaron Issac]

anybody???