How can I predict label of train images by a classifier which was trained using treebagger?

Dear All,

I have ten 3D images as train data and one image as test data. It is a classification problem. After creating feature matrix using 10 train images, I trained classifier as follow:

 B = TreeBagger(Num_Tree , Feature_A_Norm , Label_A , ...
                        'Method' , 'classification' , 'OOBPredictorImportance' , ...
                        'on' , 'MinLeafSize' , Leaf_Size , 'Prior' , 'Uniform' , ...
                        'MaxNumSplits' , Tree_Depth);

where 'Feature_A_Norm' and 'Label_A' are feature matrix and label of train images, respectively.

Now I need to predict not only label of test image but also train images.

For Test image I used:

 [~ , score_Target] = predict(B , Feature_Target_Norm);

where 'Feature_Target_Norm' is feature matrix of test image.

But I am not sure for train images should I use 'predict' function or 'oobPredict' function?I would be really appreciated if you could help me.