We examined how face recognition came to today in the series of Face Recognition History articles.
In this series of articles, we’ll study face recognition techniques and examine advantages and disadvantages against each other.
Traditional face recognition algorithms detects a person’s face image and notes some definitive features and defines them. For example the algorithm analyses face elements’ relative positions, sizes and shapes of eye, nose, chin, jaw and cheek. Then compares with the other images.
Some other algorithms standartises face images and compress the facial data then saves only info necessary. First successful face recognition systems built on this principle.
Recognition algorithms are divided in two main kinds;
1-Geometric: Checks dividive features
2-Photometric: A statistical approac defines image in parameteres and compares.
Popular recognition algorithms contain basic part analyses. These analyses are;
*linear discriminant analysis
*Elastic bunch graph matching using Fisherface algorithm
*Hidden Markov model
*the multilinear subspace learning using tensor representation,
*the neuronal motivated dynamic link matching.