April 2017
Beginner to intermediate
250 pages
6h 28m
English
One of the hindering factors of developing face-tracking systems is the tedious and error-prone process of manually annotating a large collection of images, each with a large number of points. To ease this process for the purpose of following the work in this chapter, the publicly available MUCT dataset can be downloaded from h t t p ://w w w /m i l b o . o r g /m u c t .
The dataset consists of 3,755 face images annotated with 76 point landmarks. The subjects in the dataset vary in age and ethnicity and are captured under a number of different lighting conditions and head poses.
To use the MUCT dataset with the code in this chapter, perform the following steps:
Read now
Unlock full access