A novel face recognition method based on fusion of LBP and HOG

Abstract

As one of the hot topics in the field of computer vision research, face recognition technology has received significant attention due to its potentiality for a wide range of applications in government as well as commercial purposes. In practical applications, although several existing face recognition methods have achieved good performances in specific scenes, they easily suffer from a sharp decline in recognition rate if affected by different conditions of light, expression, posture and occlusion. Among many factors, influences of complex illuminations on face recognition are particularly significant. To further improve the performance of the existing local binary pattern (LBP) operator, neighbourhood weighted average LBP (NWALBP) is first proposed for fully considering the strong correlations between pixel pairs in the neighbourhood, which extends the traditional LBP uni-layer neighbourhood template window to the bi-layer neighbourhood template window and calculates the weighted average of bi-layer neighbourhood pixels in each direction. Then, inspired by center symmetric LBP (CS-LBP), centre symmetric NWALBP (CS-NWALBP) is further proposed, which can effectively reduce computation complexity by only comparing the weighted average values of the neighbourhood pixels that are symmetric about the centre pixel. Finally, by combining the merit of histogram of oriented gradient (HOG), a feature fusion algorithm named CS-NWALBP+HOG is suggested. Several experiments have eventually demonstrated that our proposed algorithms have more robust performance under complex illumination conditions if compared with many other latest algorithms.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Small insulator target detection based on multi‐feature fusion

“…Similar research to this paper includes [15, 17, 31]. In [15], a feature extraction algorithm based on the HOG and LBP (Local Binary Pattern) algorithms and the CS‐NWALBP algorithm was proposed, which solved the sensitivity of LBP to noise and reduced the complexity of the algorithm. The effectiveness of this algorithm was experimentally verified on multiple datasets.…”

“…In the past 20 years, it is one of the most influential machine learning algorithms [25][26][27]. Equation (15) shows the original problem of SVM: In practice, SVM optimizes the classification problem by error penalty factor C and kernel function, as shown in Equation (16). The error penalty parameter C changes the learning complexity and tolerance of misclassification of the learning machine in the sample feature space so that the learning machine can find the largest soft gap surface when it seeks the largest hard gap against a wall.…”

“…Methods based solely on the local characteristics of the insulator ignore the relationships between the local features, thus losing some of the global features. With the development of multi-feature fusion technology [15][16][17], it has become possible to resolve the contradiction between the overall and local information features of images.…”

Construction of a smart face recognition model for university libraries based on FaceNet-MMAR algorithm

“…Chen T et al first proposed a feature fusion algorithm called centrosymmetric local binary pattern gradient direction histogram. Compared to other recent algorithms, this algorithm can efficiently and accurately recognize faces even in complex lighting conditions [ 8 ]. Paul K C and Aslan S proposed an optimized real-time facial recognition system to enhance AI facial recognition with an accuracy of 60.60% and 95% at 15 and 45 pixels, respectively [ 9 ].…”

Automatic Geographic Information Classification Method Based on SVM for Site Selection of Stretch Field

“…Since SVM has high recognition accuracy even for small sample training results, it avoids a large number of manually labeled image categories, and, the fast-training speed saves a lot of time and improves efficiency [11,12]. In addition, HOG is used as a training feature, and as a gradient feature, HOG features themselves have high discriminative properties [13,14]. Therefore, the use of SVM combined with HOG for training and recognition can eventually lead to accurate geographic image classification results.…”