A New Local Descriptor Based on Strings for Face Recognition

Algorithm 1

Building dictionary.

Input: Train image (matrix)
Output: Dictionary of Visual Words (DoVW)
(1)	Divide the face image into subregions.
(2)	Locate the pixels in each subregion.
(3)	Bring out the strings formed only by the four letters (R, L, U and D).
(4)	Construction of the dictionary D in each subregion i as follows:
(1)
	Where is the number of subregions, and is the number of words in each subregion, stands for the jth word from the ith subregion, following the path of successive maxima, and stands for the jth word from the ith subregion, following the path of successive minima.
(5)	Concatenate regional dictionaries into one: (2)
	Once the DoVW is created, the words are extracted from the face image by the process pursued in the next part.