New answers tagged

0 votes

How to instruct Mask RCNN to identify objects too close to each other?

Mask R-CNN should work well at this task. Before training, make sure to draw a mask on every chip in the image (put each mask in a different file). If you mask out only a small subset of the chips ...
user avatar
  • 671
1 vote

How to generate new data given a trained VAE - sample from the learned latent space or from multivariate Gaussian?

Maybe there's a misconception, we are not learning a mu and log_var but a mapping (encoder) from an image to ...
user avatar
  • 136
0 votes

How to generate new data given a trained VAE - sample from the learned latent space or from multivariate Gaussian?

I think method 1 will provide the best output. Approximating the empirical distribution of $z$ should provide decoder inputs in the subset of latent space that the decoder was trained on. Sampling ...
user avatar

Top 50 recent answers are included