One of the properties of GANs is not only generating new images from latent vectors, but the ability to manipulate the latent vectors. One example of this is to shift between two generated images. The code requires the user to manually identify the ID of the two latent vectors they wish to shift between, and the number of steps they would like to shift. Example images shows ten steps between two generated fibroblasts. This code amends the example found at: https://ch.mathworks.com/help/deeplearning/ug/train-conditional-generative-adversarial-network.html New code starts on line 253
Function version also attached
Michael Keeling (2022). Shift bewteen outputs of Conditional GAN (https://www.mathworks.com/matlabcentral/fileexchange/77813-shift-bewteen-outputs-of-conditional-gan), MATLAB Central File Exchange. 검색됨 .