Conference Paper

Face Recognition using Layered Linear Discriminant Analysis and Small Subspace

DOI: 10.1109/CIT.2010.252 Conference: 10th IEEE International Conference on Computer and Information Technology, CIT 2010, Bradford, West Yorkshire, UK, June 29-July 1, 2010
Source: DBLP

ABSTRACT

Face recognition has great demands in human recognition and recently it becomes one of the most important research areas of biometrics. In this paper, we present a novel layered face recognition method based on Fisher's linear discriminant analysis. The basic aim is to decrease FAR by reducing the face dataset to small size by applying layered linear discriminant analysis. Although, the computational complexity at the time of recognition is much higher than conventional PCA and LDA due to the weights computation for small subspace at the time of recognition, but on the other hand the layered LDA provides significant performance gain especially on similar face database. Layered LDA is insensitive to large dataset and also small sample size and it provides 93% accuracy on BANCA face database. Experimental and simulation results show that the proposed scheme has encouraging results for a practical face recognition system.

  • Source
    • "However, in practical applications of LDA, when sample dimension is greater than or close to the number of samples, the withinclass scattering matrix is not reversible. Meanwhile, it is difficult to calculate the matrix directly, which is the so-called problem of í µí± í µí±ší µí±Ží µí±™í µí±™í µí± í µí±Ží µí±ší µí±í µí±™í µí±’í µí± í µí±–í µí± §í µí±’(í µí±†í µí±†í µí±†) [21]. Therefore, we take advantage of the best transformation matrix í µí°º * to overcome the í µí±†í µí±†í µí±† problem, which is defined as follows. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Mobile sensor networks (MSNs), consisting of mobile nodes, are sensitive to network attacks. Intrusion detection system (IDS) is a kind of active network security technology to protect network from attacks. In the data gathering phase of IDS, due to the high-dimension data collected in multidimension space, great pressure has been put on the subsequent data analysis and response phase. Therefore, traditional methods for intrusion detection can no longer be applicable in MSNs. To improve the performance of data analysis, we apply K -means algorithm to high-dimension data clustering analysis. Thus, an improved K -means clustering algorithm based on linear discriminant analysis (LDA) is proposed, called LKM algorithm. In this algorithm, we firstly apply the dimension reduction of LDA to divide the high-dimension data set into 2-dimension data set; then we use K -means algorithm for clustering analysis of the dimension-reduced data. Simulation results show that LKM algorithm shortens the sample feature extraction time and improves the accuracy of K -means clustering algorithm, both of which prove that LKM algorithm enhances the performance of high-dimension data analysis and the abnormal detection rate of IDS in MSNs.
    Full-text · Article · Oct 2015 · International Journal of Distributed Sensor Networks
  • Source
    • "Two issues in face recognition algorithms are: feature representation and classification based on features. This paper presents issues involved in small data set development using L-LDA [13] by using additional clue to extract the small dataset and also resolve the conflict between two close classes. classification The proposed technique is intensive to large dataset and small subspace (SSS) problems by optimizing the seperatiablity criteria through extracting the small size dataset from large dataset. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Face recognition has a great demands in human authentication and it becomes one of the most intensive field of biometrics research areas. In this paper, we present a bio-inspired face recognition system based on linear discriminant analysis and external clue i.e. geometrical features. The use of external clue helps to identify the face among very close match and secondly it also helps in the creation of small data set. The proposed approach is insensitive to large dataset and small sample size (SSS) and it provides 94.5% accuracy on BANCA face database. Experimental and simulation results shows that the proposed scheme has encouraging results for a practical face recognition system. The computational complexity of proposed system is more than conventional LDA due to the computation of weights during recognition and in external clue but on the other it provides significant performance gain especially on similar face database.
    Preview · Conference Paper · Oct 2010
  • [Show abstract] [Hide abstract]
    ABSTRACT: Security can be enhanced through wireless sensor network using contactless biometrics and it remains a challenging and demanding task due to several limitations of wireless sensor network. Network life time is very less if it involves image processing task due to heavy energy required for image processing and image communication. Contactless biometrics such as face recognition is most suitable and applicable for wireless sensor network. Distributed face recognition in WSN not only help to reduce the communication overload but it also increase the node life time by distributing the work load on the nodes. This paper presents state-of-art of biometrics in wireless sensor network.
    No preview · Conference Paper · Jan 2010
Show more