Details of code

[Aashima-Gupta]

Hi, Your code really helped me. However, I have a doubt
What does this part of the do in the Train_Images.py file?

with open(label_name_map_file, 'wb') as handle:
                pickle.dump(label_name_map, handle, protocol=pickle.HIGHEST_PROTOCOL) 
#Dump the label:name map

[Aashima-Gupta]

Also tell what does this code do:

 person_name = os.path.split(i)[1].split(".")[0]
                        if name2label_map.has_key(person_name):
                                name = name2label_map[person_name]
                        else:
                                count += 1
                                name2label_map[person_name] = count
                                name = count
                                label2name_map[count] = person_name

[GokulNC]

1. It's used to dump the label to names map as a file.
   Essentially, this is loaded by the Recognizer.py, since OpenCV returns only returns labels for each face (for which we've to take responsibility of finding the right name)

2. If you checked the images_db folder, the files will be in the format person_name.<anything>.png
   For each image in that folder, the Train_Images.py loads it and saves the name to label map.

For example, for a file named Aashima.1.png, it might stored the name Aashima to the label 1 and label 1 for Aashima. Say for another file Gupta.53.png it might store the name Gupta to the label 2 and so on.