05-05-2012, 03:15 PM
Face Detection using Neural Networks
phil.pdf (Size: 1.41 MB / Downloads: 257)
Introduction
Pattern recognition is a modern day machine intelligence problem with numerous
applications in a wide field, including Face recognition, Character recognition,
Speech recognition as well as other types of object recognition. The field of pattern
recognition is still very much in it is infancy, although in recent years some of the
barriers that hampered such automated pattern recognition systems have been lifted
due to advances in computer hardware providing machines capable of faster and
more complex computation.
Face Detection Approaches – An Overview
s and Low [1] conducted a survey on face detection techniques, and identified
two broad categories that separate the various approaches, appropriately named
feature-based and image-based approaches. Each category will be explained, and
the work completed will be presented, providing a brief yet thorough overview of the
various face detection techniques.
Feature-Based Approach
Hjelm
s and Low [1] divide the group of feature-based systems into three subcategories:
low-level analysis, feature analysis, and active shape models.
Low-level Analysis
Low-level Analysis deals with the segmentation of visual features using the various
properties of the pixels, predominantly gray-scale or colour.
Edge representation (detecting changes in pixel properties) was first implemented by
Sakai et al [2] for detecting facial features in line drawings. Craw et al [3] developed
this further to trace a human head outline, allowing feature analysis to be constrained
to within the head outline. Various operators are used to detect the presence of an
edge, including the Sobel operator, the Marr-Hildreth operator, and a variety of first
and second derivatives of Gaussians.
Feature Analysis
Low-level analysis introduces ambiguity which can be solved by high level feature
analysis, often through the use of some additional knowledge about the face. There
are two approaches for the application of this additional knowledge (commonly face
geometry).