Deep Learning vs Object Separation CT Scan Analysis Explained

https://rigaku.com/
[47:34-49:20]
object separation is a process to separate in this example for example this is a sandstone CT scan You see the sand grains When you just do segmentation you are just labeling those pixels or voxels as sand Then when you do object separation you're gonna separate individual sand grains Then you will index them so that you can analyze the grain size Then you see the histogram of the grain size It's color-coded and you can see most of them are pretty small but you can easily identify green or red grains that are large So to do this type of analysis you have to do object separation after the simple segmentation And once you do this you can analyze either the size of the grains or maybe orientation of fibers or aspect ratio of fibers Anything you can quantitatively measure on each object you can analyze So that's object separation The other topic we didn't cover much is deep learning Now I just wanted to quickly show you what deep learning can do on internal segmentation So this is a CT scan a cross-section of a tiny particle inside of an orally disint- disintegrating tablet and this tiny particle has multiple coatings Let's say we want to segment those coatings to do some quantitative analysis and if you use just simple thresholding this is the best that you can get

Видео Deep Learning vs Object Separation CT Scan Analysis Explained канала Rigaku Corporation