how can I give input as image in GAN example to generate new images ?

I m following this mathworks example ,

https://in.mathworks.com/help/deeplearning/ug/train-conditional-generative-adversarial-network.html

Trained the network on my data set . But now I want to give image as input and generate new images . In this example it takes random valued and generate images for selected class .

I tried to resize image to 10x10 and than reshape to [1,1,100] that is equal to the input of generator. But got very bas results .

I am thinking to retarin with change size of input in genrator [1,1,4096] ( 64*64 size of image image ) . Or any other Idea ?

thank you

Answer 1

[Nagwa megahed]

you can read this Monitor GAN Training Progress and Identify Common Failure Modes