Local Derivative Pattern Versus Local Binary Pattern: Face Recognition With High-Order Local Pattern Descriptor

• Published in: IEEE transactions on image processing Vol. 19; no. 2; pp. 533 - 544
• Main Authors: Zhang, Baochang, Gao, Yongsheng, Zhao, Sanqiang, Liu, Jianzhuang
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.02.2010; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Applied sciences; Biometric Identification - methods; Coding, codes; Data mining; Databases, Factual; Derivatives; Detection, estimation, filtering, equalization, prediction; Encoding; Exact sciences and technology; Face - anatomy & histology; Face recognition; Feature based; Feature extraction; Gabor feature; high-order local pattern; Humans; Image databases; Image processing; Information, signal and communications theory; Linear discriminant analysis; local binary pattern (LBP); local derivative pattern (LDP); Pattern recognition; Performance evaluation; Principal component analysis; Signal and communications theory; Signal processing; Signal, noise; Spatial databases; Telecommunications and information theory; Biometrics; Performance evaluation; Gabor filter; Grey level image; Face recognition; Image processing; high-order local pattern; Pattern recognition; Gabor feature; N order; Coding; First order; Database; local binary pattern (LBP); local derivative pattern (LDP); Automatic recognition
• ISSN: 1057-7149; 1941-0042; 1941-0042
• DOI: 10.1109/TIP.2009.2035882

This paper proposes a novel high-order local pattern descriptor, local derivative pattern (LDP), for face recognition. LDP is a general framework to encode directional pattern features based on local derivative variations. The nth -order LDP is proposed to encode the ( n -1) th -order local derivative direction variations, which can capture more detailed information than the first-order local pattern used in local binary pattern (LBP). Different from LBP encoding the relationship between the central point and its neighbors, the LDP templates extract high-order local information by encoding various distinctive spatial relationships contained in a given local region. Both gray-level images and Gabor feature images are used to evaluate the comparative performances of LDP and LBP. Extensive experimental results on FERET, CAS-PEAL, CMU-PIE, Extended Yale B, and FRGC databases show that the high-order LDP consistently performs much better than LBP for both face identification and face verification under various conditions.