Title :
Face recognition using eyes, nostrils and mouth features
Author :
Paul, Sushil Kumar ; Uddin, Mohammad Shorif ; Bouakaz, Saida
Author_Institution :
Dept. of Comput. Sci. & Eng., Jahangirnagar Univ., Dhaka, Bangladesh
Abstract :
This paper describes a face recognition algorithm that extracts the eyes, nostrils and mouth features from cumulative distribution function (CDF) by applying Otsu thresholding. The algorithm, which is inspired by the probability of white pixels of binary facial image, has been tested using the BioID frontal face large database in different illuminations, expressions and lighting conditions. Illumination and lighting variations are addressed using a selective equalization technique. The experimental results have confirmed an average recognition rate of 93.55%.
Keywords :
face recognition; feature extraction; image resolution; image segmentation; statistical distributions; BioID frontal face; CDF; Otsu thresholding; binary facial image white pixel probability; cumulative distribution function; expression conditions; eye feature extraction; face recognition algorithm; illumination conditions; lighting conditions; mouth feature extraction; nostril feature extraction; selective equalization technique; Databases; Face; Face recognition; Facial features; Feature extraction; Lighting; Mouth; Cumulative distribution function; Otsu thresolding; face recognition; facial features; probability of white pixels;
Conference_Titel :
Computer and Information Technology (ICCIT), 2013 16th International Conference on
Conference_Location :
Khulna
DOI :
10.1109/ICCITechn.2014.6997378
