Tiankang Xie

We can optimize the function here

1

https://github.com/cosanlab/nltools/blob/a2a54df71972f41e4efc77dd861fd07bd214f0ec/nltools/data/brain_data.py#L958

In the current version, the landmark is scaled down to the original image size, but not the desired output size as set by the user. Which way shall we decide?

Currently the pytorch read_image only accepts jpg & png but not other file extensions. If we write our own custom image loader it will enable more file extensions. This may...

Bug fix for enlarging bounding boxes that is beyond the image boundary

1

Problem: sometimes bounding boxes are beyond image boundary. Some bugs (negative values) were seen during Rescale processes. Details: Calling detector.process_image on one of the CK+ images (original size 200x200) raises...

Just wanted to share with you and the lab about this useful python toolbox on audio sync: https://github.com/benfmiller/audalign Very effective on my FNL alignment tasks. ``` import audalign as ad...