Showing papers in "Pattern Recognition Letters in 2005"

TL;DR: This work designs classifiers based on the well-known fisherface method and demonstrates that the proposed method comes with better performance when compared with other template-based techniques and shows substantial insensitivity to large variation in light direction and facial expression.

TL;DR: An innovative algorithm named 2D-LDA is proposed, which directly extracts the proper features from image matrices based on Fisher's Linear Discriminant Analysis, and achieves the best performance.

TL;DR: The results of comparative study show that the proposed PCAES index has high ability in producing a good cluster number estimate and in addition, it provides a new point of view for cluster validity in a noisy environment.

TL;DR: The axiom definition of distance measure between intuitionistic fuzzy sets (IFSs) is introduced and some distance measures are proposed and corresponding proofs are given.

TL;DR: This work presents a system for online handwritten signature verification, approaching the problem as a two-class pattern recognition problem, and received the first place at SVC2004 with a 2.8% error rate.

TL;DR: A real time and robust method of license plate location that first extracts out the vertical edges of the car image using image enhancement and Sobel operator, then removes most of the background and noise edges by an effective algorithm, and finally searches the plate region by a rectangle window in the residual edge image.

TL;DR: In this paper, a singular value decomposition (SVD)-based watermarking scheme is proposed, which preserves both one-way and non-symmetric properties, usually not obtainable in DCT and DFT transformations.

TL;DR: A new LDA method is proposed that attempts to address the SSS problem using a regularized Fisher's separability criterion and a scheme of expanding the representational capacity of face database is introduced to overcome the limitation that the LDA-based algorithms require at least two samples per class available for learning.

TL;DR: This work designs an automatic SRG algorithm, along with a boundary-oriented parallel pixel labeling technique and an automatic seed selection method, and a seed tracking algorithm is proposed for automatic moving object extraction.

TL;DR: A wavelet based logo-watermarking scheme for copyright protection of digital image using a visually meaningful gray scale logo as watermark is presented and is robust to wide variety of attacks.

TL;DR: It is shown that when two motion-blurred images are available, having different blur directions, image restoration can be improved substantially and the direction of the motion blur and the PSF (Point Spread Function) of the blur can be computed robustly.

TL;DR: This paper proposes a method which combines the methods of the interpretive structural modeling and the analytic network process (ANP) procedures to deal with the problem of the subsystems' interdependence and feedback.

TL;DR: This paper presents an analysis of design principles implicitly used in defining cluster validity indices and reviews a variety of existing clusters validity indices in the light of these principles, and proposes six new indices.

TL;DR: It is shown that the fuzzy-rough set attribute reduction algorithm is not convergent on many real datasets due to its poorly designed termination criteria; and the computational complexity of the algorithm increases exponentially with increase in the number of input variables and in multiplication with the size of data patterns.

TL;DR: It is shown that both the writer identification and the writer verification tasks can be carried out using local features such as graphemes extracted from the segmentation of cursive handwriting, making the approach general and very promising for large scale applications in the domain of handwritten document querying and writer verification.

TL;DR: The algorithms shown in this paper generalize and extend the concept of Bounded Partial Correlation, previously devised for a template matching process based on the Normalized Cross-Correlation function (NCC), to allow to skip rapidly most of the expensive calculations required to evaluate the ZNCC at those image points that cannot improve the best correlation score.

TL;DR: A comparison of the two classifiers in off-line signature verification using random, simple and simulated forgeries to observe the capability of the classifiers to absorb intrapersonal variability and highlight interpersonal similarity.

TL;DR: Two hierarchical IDS frameworks using Radial Basis Functions (RBF) are proposed and can detect network intrusions in real-time, train new classifiers for novel intrusions automatically, and modify their structures adaptively after new classifier are trained.

TL;DR: This paper presents a simple index called list of clusters (LC), which is shown to require little space, to be suitable both for main and secondary memory implementations, and most importantly to be very resistant to the intrinsic dimensionality of the data set.

TL;DR: Experimental results show that this method for automatic road extraction in rural and semi-urban areas is efficient in extracting and defining road networks from high resolution satellite and aerial imagery.

TL;DR: Methods of selecting the appropriate granule size and efficient computation of rough entropy are described, which results in minimization of roughness in both object and background regions; thereby determining the threshold of partitioning.

TL;DR: A hybrid optimization scheme for multiple thresholding by the criteria of Otsu's minimum within-group variance and Gaussian function fitting, which shows that the NM-PSO-Otsu could expedite the OTSu's method efficiently to a great extent in the case of multi-level thresholding, and that theNM- PSO-curve method could provide better effectiveness in the context of visualization, object size and image contrast.

TL;DR: Experimental results show that the simple features like X-, Y-coordinates, the speed of writing and the angle with the X-axis are amongst the most consistent.

TL;DR: The proposed image thresholding approach based on the index of entropy maximization of the 2-D grayscale histogram is proposed to deal with infrared image and can get ideal segmentation result with less computation cost.

TL;DR: It is proved that the principal component analysis (PCA) and the linear discriminant analysis (LDA) can be directly implemented in the discrete cosine transform (DCT) domain and the results are exactly the same as the one obtained from the spatial domain.

TL;DR: A conditional probability based, efficient method to extract the significant attributes from a database and shows how the classification methodology can be used for cost-sensitive learning where both accuracy and precision of prediction are important.

TL;DR: This paper presents a modified FCM algorithm for bias estimation and segmentation of MRI, formulated by modifying the objective function of the standard FCM and it has the advantage that it can be applied at an early stage in an automated data analysis before a tissue model is available.

TL;DR: The results show that, compared with MHD, the new Hausdorff distance can dispose of the noisy image matching in a more desirable manner, due to the fact that the comprehensive reflection of the gray information of neighbor pixels in the determination of the Hausdorf distance is taken into account.

TL;DR: The results show that stratified evolutionary prototype selection consistently outperforms the non-evolutionary ones, the main advantages being: better instance reduction rates, higher classification accuracy and reduction in resources consumption.

TL;DR: A general locally adaptive thresholding method using neighborhood processing is presented that provides superior image segmentation to existing thresholding methods for images that are severely degraded due to noise, poor illumination and shadow.