I'm not aware of any implementation with these features, but they are both on the roadmap for the linked repo. Both should also be achievable. Batch processing,

I'm not aware of any implementation with these features, but they are both on the roadmap for the linked repo. Both should also be achievable. Batch processing, in particular, will be a straight-forward change and should result in quite a speed-up. Although it will require the input images to have the same dimensions.