The source of target variables to train Tissue Classifier

[ShangWeiKuo]

Excuse me.

I've tried to train tissue classifier, and I've realized how predictor variables are generated. However, I have no idea about the data format and the source of target variables.

Could you explain the data format and the source of target variables?

[unimusic]

Hi, can you again explain exactly what you are doing with an example? Thanks.

[ShangWeiKuo]

According to the picture below, the predictor variables of tissue classifier are the registered scans whose variables are named "image".





If I would like to retrain tissue classifier, I have to implement "clf.fit(X, y)".

Now I've known that "image" is predictor variable "X" in segUtils.py, but what is target variable "y" ?

[unimusic]

Target variable y is the predicted class of each voxel from x_predict, which is the flattened image.

[ShangWeiKuo]

I know that y is the predicted class of each voxel from x_predict. However, How could I get the target variable y from training dataset instead of the predicted one if I would like to retrain tissue classifier?

[unimusic]

Ah, I didn't provide that in this repo. The clf is pretrained. You can create your own new predictor if you wish.