Cost-Sensitive Sparsity Preserving Projections for Face Recognition

2013 ◽  
Vol 760-762 ◽  
pp. 1615-1620 ◽  
Author(s):  
Xiao Yuan Jing ◽  
Wen Qian Li ◽  
Hao Gao ◽  
Yong Fang Yao ◽  
Jiang Yue Man
Keyword(s):  
Face Recognition ◽  
Sparse Representation ◽  
Dimensional Space ◽  
Recognition Rate ◽  
Recognition Task ◽  
Cost Information ◽  
The Public ◽  
Sparsity Preserving Projection ◽  
Low Dimensional ◽  
The Cost

As one of the most popular research topics, sparse representation (SR) technique has been successfully employed to solve face recognition task. Though current SR based methods prove to achieve high classification accuracy, they implicitly assume that the losses of all misclassifications are the same. However, in many real-world face recognition applications, this assumption may not hold as different misclassifications could lead to different losses. Driven by this concern, in this paper, we propose a cost-sensitive sparsity preserving projections (CSSPP) for face recognition. CSSPP considers the cost information of sparse representation while calculating the sparse structure of the training set. Then, CSSPP employs the sparsity preserving projection method to achieve the projection transform and keeps the sparse structure in the low-dimensional space. Experimental results on the public AR and FRGC face databases are presented to demonstrate that both of the proposed approaches can achieve high recognition rate and low misclassification loss, which validate the efficacy of the proposed approach.

Online Long-Term Object Tracking Based on Compressed Haar-Like Features via Sparse Representation

2014 ◽  
Vol 989-994 ◽  
pp. 1610-1614
Author(s):  
Ming Zhao ◽  
Lu Ping Wang ◽  
Lu Ping Zhang
Keyword(s):  
Sparse Representation ◽  
Target Location ◽  
Model Updating ◽  
Dimensional Space ◽  
Reconstruction Error ◽  
Illumination Changes ◽  
Spatio Temporal ◽  
Low Dimensional ◽  
Scale Variation

Online long-term tracking is a challenging problem as data streams change over time. In this paper, sparse representation has been applied to visual tracking by finding the most correct sample with minimal reconstruction error using compressed Haar-like features. However, most sparse representation tracking algorithm introduce l1 regularization into the PCA reconstruction using samples directly, which leads to complexity computation and can not adapt to occlusion, rotation and change in size. Our model updating not only uses the samples from the training set, but also generates the warped versions (include scale variation, rotation, occlusion and illumination changes) for the previous tracking result. Also, we do not use the samples in models for sparse representation directly, but the Haar-like features instead which are compressed in a very low-dimensional space. In addition, we use a robust and fast algorithm which exploits the spatio-temporal context for predicting the target location in the next frame. This step will lead to the reduction of the searching range by the detector. We demonstrate the proposed method is able to track objects well under pose and scale variation, rotation, occlusion and illumination with great real-time performance on challenging image sequences.

Research of Feature Extraction Method on Facial Expression Change

2011 ◽  
Vol 211-212 ◽  
pp. 813-817 ◽  
Author(s):  
Jin Qing Liu ◽  
Qun Zhen Fan
Keyword(s):  
Feature Extraction ◽  
Wavelet Transform ◽  
Facial Expression ◽  
Extraction Method ◽  
Dimensional Space ◽  
Recognition Rate ◽  
Low Frequency ◽  
Expression Change ◽  
Feature Extraction Method ◽  
Low Dimensional

In this paper, the purpose is to find a method that can be more suited to facial expression change and also improve the recognition rate. The proposed system contains three parts, wavelet transform, Fisher linear discriminant method feature extraction and face classification. The basic idea of the proposed method is that first extract the low-frequency components through wavelet transform, then the low-frequency images mapped into a low-dimensional space by PCA transform, and finally the utilization of LDA feature extraction method in low-dimensional space. The algorithms were tested on ORL and Yale face database, respectively. Experimental results shows that the proposed method not only improve the recognition rate, but also improve the recognition speed. This method can effectively overcome the impact of expression changes on face recognition, and play a certain role in inhibition of expression.

scholarly journals Locally Linear Discriminate Embedding for Face Recognition

2009 ◽  
Vol 2009 ◽  
pp. 1-8 ◽  
Author(s):  
Eimad E. Abusham ◽  
E. K. Wong
Keyword(s):  
Face Recognition ◽  
Dimensional Space ◽  
Linear Structure ◽  
Data Representation ◽  
High Dimensional ◽  
Fast Processing ◽  
Novel Method ◽  
Computational Simplicity ◽  
Low Dimensional ◽  
Locally Linear

A novel method based on the local nonlinear mapping is presented in this research. The method is called Locally Linear Discriminate Embedding (LLDE). LLDE preserves a local linear structure of a high-dimensional space and obtains a compact data representation as accurately as possible in embedding space (low dimensional) before recognition. For computational simplicity and fast processing, Radial Basis Function (RBF) classifier is integrated with the LLDE. RBF classifier is carried out onto low-dimensional embedding with reference to the variance of the data. To validate the proposed method, CMU-PIE database has been used and experiments conducted in this research revealed the efficiency of the proposed methods in face recognition, as compared to the linear and non-linear approaches.

Study on dual peg-in-hole insertion using of constraints formed in the environment

2017 ◽  
Vol 44 (6) ◽  
pp. 730-740 ◽  
Author(s):  
Jianhua Su ◽  
Rui Li ◽  
Hong Qiao ◽  
Jing Xu ◽  
Qinglin Ai ◽  
...  
Keyword(s):  
Configuration Space ◽  
Dimensional Space ◽  
Three Dimensional ◽  
Force Sensor ◽  
High Dimensional ◽  
Content Type ◽  
Attractive Region ◽  
Low Dimensional ◽  
The Cost ◽  
Yaw Angle

Purpose The purpose of this paper is to develop a dual peg-in-hole insertion strategy. Dual peg-in-hole insertion is the most common task in manufacturing. Most of the previous work develop the insertion strategy in a two- or three-dimensional space, in which they suppose the initial yaw angle is zero and only concern the roll and pitch angles. However, in some case, the yaw angle could not be ignored due to the pose uncertainty of the peg on the gripper. Therefore, there is a need to design the insertion strategy in a higher-dimensional configuration space. Design/methodology/approach In this paper, the authors handle the insertion problem by converting it into several sub-problems based on the attractive region formed by the constraints. The existence of the attractive region in the high-dimensional configuration space is first discussed. Then, the construction of the high-dimensional attractive region with its sub-attractive region in the low-dimensional space is proposed. Therefore, the robotic insertion strategy can be designed in the subspace to eliminate some uncertainties between the dual pegs and dual holes. Findings Dual peg-in-hole insertion is realized without using of force sensors. The proposed strategy is also used to demonstrate the precision dual peg-in-hole insertion, where the clearance between the dual-peg and dual-hole is about 0.02 mm. Practical implications The sensor-less insertion strategy will not increase the cost of the assembly system and also can be used in the dual peg-in-hole insertion. Originality/value The theoretical and experimental analyses for dual peg-in-hole insertion are proposed without using of force sensor.

An Algorithm of Frequent Neighboring Class Set Mining with Constraint Class

2011 ◽  
Vol 211-212 ◽  
pp. 808-812
Author(s):  
Gang Fang ◽  
Hong Ying ◽  
Jiang Xiong ◽  
Yuan Bin Wu
Keyword(s):  
Feature Extraction ◽  
Wavelet Transform ◽  
Dimensional Space ◽  
Recognition Rate ◽  
Low Frequency ◽  
Expression Change ◽  
Linear Discriminant ◽  
Feature Extraction Method ◽  
Low Dimensional ◽  
The Impact

In this paper, the purpose is to find a method that can be more suited to facial expression change and also improve the recognition rate. The proposed system contains three parts, wavelet transform, Fisher linear discriminant method feature extraction and face classification. The basic idea of the proposed method is that first extract the low-frequency components through wavelet transform, then the low-frequency images mapped into a low-dimensional space by PCA transform, and finally the utilization of LDA feature extraction method in low-dimensional space. The algorithms were tested on ORL and Yale face database, respectively. Experimental results shows that the proposed method not only improve the recognition rate, but also improve the recognition speed. This method can effectively overcome the impact of expression changes on face recognition, and play a certain role in inhibition of expression.

Face Recognition Using 2D-FLDA Based on Approximate SVD Obtained with Kronecker Products with One Training Sample

2019 ◽  
pp. 1-9
Author(s):  
Ümit Çiğdem Turhal
Keyword(s):  
Face Recognition ◽  
Kronecker Product ◽  
Dimensional Space ◽  
Recognition Performance ◽  
Recognition Task ◽  
Linear Structure ◽  
Training Sample ◽  
Kronecker Products ◽  
Linear Discriminant ◽  
Lower Dimensional

Aims: In a face recognition task, it is a challenging problem to find lots of images for a person. Even, sometimes there can be only one image, available for a person. In these cases many of the methods are exposed to serious performance drops even some of these fail to work. Recently this problem has become remarkable for researchers. In some of these studies the database is extended using a synthesized image which is constructed from the singular value decomposition (SVD) of the single training image. In this paper, for such a method, SVD based 2 Dimensional Fisher Linear Discriminant Analysis (2D-FLDA), it is proposed a new approach to find the SVD of the image matrix with the aim of to increase the recognition performance. Study Design: In this paper, in a face recognition task with 2D-FLDA, in one training sample case, instead of original SVD of the image matrix, the approximate SVD of its based on multiple kronecker product sums is used. In order to obtain it, image matrix is first reshaped thus it is to be lower dimensional matrices and, then the sum of multiple kronecker products (MKPS) is applied in this lower dimensional space. Methodology: Experiments are performed on two known databases Ar-Face and ORL face databases. The performance of the proposed method is evaluated when there are facial expression, lightning conditions and pose variations. Results: In each experiment, the approximate SVD approach based on multiple kronecker product sum gets approximately 3% better results when compared with the original SVD. Conclusion: Experimental results verify that the proposed method achieves better recognition performance over the traditional one. The reason for this is the proposed approximate SVD has the advantages of simplicity, and also as the kronecker factors possess additional linear structure, kronecker product can capture potential self-similarity.

ADAPTIVE CLASSIFIER INTEGRATION FOR INVARIANT FACE RECOGNITION (ACIIFR)

2002 ◽  
Vol 16 (06) ◽  
pp. 749-772 ◽  
Author(s):  
G. A. KHUWAJA ◽  
M. S. LAGHARI
Keyword(s):  
Face Recognition ◽  
Network Architecture ◽  
Recognition Rate ◽  
Recognition Task ◽  
Recognition System ◽  
Face Image ◽  
Divide And Conquer ◽  
Accuracy Rate ◽  
Learning Time ◽  
The Face

The integration of multiple classifiers promises higher classification accuracy and robustness than can be obtained with a single classifier. We address two problems: (a) automatic recognition of human faces using a novel fusion approach based on an adaptive LVQ network architecture, and (b) improve the face recognition up to 100% while maintaining the learning time per face image constant, which is an scalability issue. The learning time per face image of the recognition system remains constant irrespective of the data size. The integration of the system incorporates the "divide and conquer" modularity principles, i.e. divide the learning data into small modules, train individual modules separately using compact LVQ model structure and still encompass all the variance, and fuse trained modules to achieve recognition rate nearly 100%. The concept of Merged Classes (MCs) is introduced to enhance the accuracy rate. The proposed integrated architecture has shown its feasibility using a collection of 1130 face images of 158 subjects from three standard databases, ORL, PICS and KU. Empirical results yield an accuracy rate of 100% on the face recognition task for 40 subjects in 0.056 seconds per image. Thus, the system has shown potential to be adopted for real time application domains.

scholarly journals ReS2 Charge Trapping Synaptic Device for Face Recognition Application

2020 ◽  
Vol 15 (1) ◽  
Author(s):  
Ze-Hui Fan ◽  
Min Zhang ◽  
Lu-Rong Gan ◽  
Lin Chen ◽  
Hao Zhu ◽  
...  
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Face Recognition ◽  
Recognition Rate ◽  
Recognition Task ◽  
Charge Trapping ◽  
Long Term Potentiation ◽  
Efficient System ◽  
Artificial Neural ◽  
Conductance States

AbstractSynaptic devices are necessary to meet the growing demand for the smarter and more efficient system. In this work, the anisotropic rhenium disulfide (ReS2) is used as a channel material to construct a synaptic device and successfully emulate the long-term potentiation/depression behavior. To demonstrate that our device can be used in a large-scale neural network system, 165 pictures from Yale Face database are selected for evaluation, of which 120 pictures are used for artificial neural network (ANN) training, and the remaining 45 pictures are used for ANN testing. A three-layer ANN containing more than 105 weights is proposed for the face recognition task. Also 120 continuous modulated conductance states are selected to replace weights in our well-trained ANN. The results show that an excellent recognition rate of 100% is achieved with only 120 conductance states, which proves a high potential of our device in the artificial neural network field.

A Sparse Representation-Based Face Recognition Using Mixed Group Sparsity

2014 ◽  
Vol 602-605 ◽  
pp. 1660-1665
Author(s):  
Ti Jian Cai ◽  
Xiao Ping Fan ◽  
Jun Xu
Keyword(s):  
Face Recognition ◽  
Sparse Representation ◽  
Recognition Rate ◽  
Search Space ◽  
Structured Sparsity ◽  
Mixed Group ◽  
Group Sparsity ◽  
Sparse Models ◽  
Length Group ◽  
The Face

Empirical evidence shows that introducing additional structured priors can reduce complexity of coding data, and achieve better performance. To improve the performance of sparse representation-based classification (SRC), the article based on the potential correlations between the elements of dictionary gets a mixed group sparsity which is composed of dynamic group sparsity and fixed-length group sparsity. To solve the structured sparsity efficiently, structured greedy algorithm (structOMP) is redesigned to fit the new structure. The modification includes search space and its neighbor. Finally, three sparse models are compared by experiments of face recognition, and the results show that the mixed group sparsity can improve the face recognition rate of other sparse models by 10% or more in dealing with corrupted data.

Face Recognition Algorithm Based on Haar-Like Features and Gentle Adaboost Feature Selection via Sparse Representation

2015 ◽  
Vol 742 ◽  
pp. 299-302 ◽  
Author(s):  
Qing Wei Wang ◽  
Zi Lu Ying ◽  
Lian Wen Huang
Keyword(s):  
Feature Selection ◽  
Face Recognition ◽  
Sparse Representation ◽  
Recognition Rate ◽  
Recognition Algorithm ◽  
Adaboost Algorithm ◽  
Face Images ◽  
Sparse Representation Classification ◽  
Feature Dimension

This paper proposed a new face recognition algorithm based on Haar-Like features and Gentle Adaboost feature selection via sparse representation. Firstly, All the images including face images and non face images are normalized to size and then Haar-Like features are extracted . The number of Haar-Like features can be as large as 12,519. In order to reduce the feature dimension and retain the most effective features for face recognition, Gentle Adaboost algorithm is used for feature selection. Selected features are used for face recognition via sparse representation classification (SRC) algorithm. Testing experiments were carried out on the AR database to test the performance of the new proposed algorithm. Compared with traditional algorithms like NS, NN, SRC, and SVM, the new algorithm achieved a better recognition rate. The effect of face recognition rate changing with feature dimension showed that the new proposed algorithm performed a higher recognition rate than SRC algorithm all the time with the increasing of feature dimension, which fully proved the effectiveness and superiority of the new proposed algorithm.

Sign in / Sign up

Export Citation Format

Share Document