Preprocessing before being input into the Keras–Python code
The VGG annotator tool generates an output CSV file that needs to be preprocessed in a format that is acceptable by our TensorFlow code. The output of the annotation is a CSV file that shows each key-point in a row format, with 16 rows per image; we need to preprocess this file so that we have one row per image. There are 33 columns indicating 32 key point values and 1 image value, shown as follows:
(x0,y0), (x1,y1), (x2, y2), ….,(x15,y15), image file name
You can use a custom Python program to convert this, although it is not shown here. The GitHub page includes the processed CSV file for your reference at https://github.com/PacktPublishing/Mastering-Computer-Vision-with-TensorFlow-2.0/blob/master/Chapter03/testimgface.csv ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access