Local gradient increasing pattern (LGIP) for facial representation and gender recognition

Authors:
Lu Bing Zhou;Han Wang
Affiliations:
School of Electrical & Electronic Engineering, Nanyang Technological University, Singapore, Singapore;School of Electrical & Electronic Engineering, Nanyang Technological University, Singapore, Singapore
Venue:
ICIAR'12 Proceedings of the 9th international conference on Image Analysis and Recognition - Volume Part II
Year:
2012

Citing 8
Cited 0

Multiresolution Gray-Scale and Rotation Invariant Texture Classification with Local Binary Patterns

IEEE Transactions on Pattern Analysis and Machine Intelligence
Gender and Ethnic Classification of Face Images

FG '98 Proceedings of the 3rd. International Conference on Face & Gesture Recognition
Face Description with Local Binary Patterns: Application to Face Recognition

IEEE Transactions on Pattern Analysis and Machine Intelligence
Evaluation of Gender Classification Methods with Automatically Detected and Aligned Faces

IEEE Transactions on Pattern Analysis and Machine Intelligence
Gender Classification Using Local Directional Pattern (LDP)

ICPR '10 Proceedings of the 2010 20th International Conference on Pattern Recognition
Combining contrast information and local binary patterns for gender classification

SCIA'11 Proceedings of the 17th Scandinavian conference on Image analysis
Gender classification based on boosting local binary pattern

ISNN'06 Proceedings of the Third international conference on Advnaces in Neural Networks - Volume Part II
The CAS-PEAL Large-Scale Chinese Face Database and Baseline Evaluations

IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans

Quantified Score

Hi-index	0.00

Visualization

Abstract

A robust facial representation is an essential component for gender classification. This paper introduces a new local feature, Local Gradient Increasing Pattern (LGIP), which expresses the local intensity increasing trend. A LGIP feature is to encode intensity increasing trends in 8 orientations at each pixel using signs of directional gradient responses, and overall increasing trend is assigned with a decimal label. A facial image is partitioned into overlapping regions from which LGIP histograms are obtained and concatenated into a single feature vector. Gender classification is carried out using SVM classifier based on the LGIP-based facial descriptor. We investigate the influence to recognition rates by two factors, image resolution and person-dependent/independent condition. Experiments are performed on two replicable image sets from CAS-PEAL and FERET databases, and the results show that our method achieves better performance than many other methods.