PhD Seminar Course on

Face Processing: from face detection to recognition

Cagliari, Jan. 20 -- Jan. 26, 2010


Prof. Sebastien Marcel     
Idiap research institute - Martigny, Switzerland


8 hours, Jan. 20 -- Jan. 26, 2010


Lecture 1 (2 hours):    Wednesday,      11:00 --- 13:00,     Jan. 20

Lecture 2 (2 hours):    Thursday,          11:00 --- 13:00,     Jan. 21

Lecture 3 (2 hours):    Monday,            11:00 --- 13:00,     Jan. 25

Lecture 4 (2 hours):    Tuesday,            11:00 --- 13:00,     Jan. 26


Mocci Classroom (DIEE, Building A)


Over the last ten years, face processing (detection and recognition) has become an important research area in computer vision, due to its wide range of possible applications such as human-computer interaction, video surveillance systems, biometrics and content-based multimedia indexing  and retrieval.

The goal of face detection is to determine whether or not there are any faces in the image and, if present, their location. It is the crucial first step of any application that involves face processing systems. Thus, accurate and fast human face detection is the key to a successful operation.
Face detection is a challenging problem because faces highly vary in size, shape, color, texture and location. Their overall appearance can also be influenced by lighting conditions, facial expression, occlusion or facial features, such as beards, mustaches and glasses. Another challenging problem comes from the orientation (upright, rotated) and the pose (frontal to profile) of the face.
The first part of these lectures will introduce the basics of face detection and will review some of the numerous techniques proposed to counter these issues.
Face recognition has existed for more than 30 years and has been particularly active since the early 1990s. Researchers from many different fields (from psychology, pattern recognition, neuroscience, computer graphic and computer vision) have attempted to create and understand face recognition systems.

This has led to many different techniques often divided into two groups:

  1. holistic matching methods
  2. feature-based matching methods. Holistic approaches use the whole face as one input while feature-based methods extract multiple features. However, recently several of the most advanced methods can be considered both feature-based (parts-based) and holistic.

The second part of these lectures will address in details the most famous and successful face recognition techniques including the recent development that make use of Local Binary Patterns.


  • Face detection and facial feature localization
  • Illumination normalization
  • Feature extraction (Discrete Cosine Transform, Gabor, Local Binary Patterns, ...) 
  • Holistic and feature-bases methods
  • Discriminative vs generative approaches
  • Performance evaluation


Ing. Gian Luca Marcialis
Dep. of Electrical and Electronic Engineering
University of Cagliari, Italy
Email: marcialis[at]