B. H. Shekar

A simple and robust line detection algorithm based on small eigenvalue analysis

In this paper, a simple and robust algorithm is proposed for detecting straight line segments in an edge image. The proposed algorithm is based on small eigenvalue analysis. The statistical and geometrical properties of the small eigenvalue of the covariance matrix of a set of edge pixels over a connected region of support are explored for the purpose of straight line detection. The approach scans an input edge image from the top left corner to the bottom right corner with a moving mask of size k × k for some odd integer k > 1. At every stage, the small eigenvalue of the covariance matrix of the edge pixels covered by the mask and connected to the center pixel of the mask is computed. These pixels are said to be linear edge pixels if the computed small eigenvalue is less than a pre-defined threshold value. Several experiments have been conducted on various images with considerable background noise and also with significant edge point location errors to reveal the efficacy of the proposed model. The results of the experiments emphasize that the proposed model outperforms other models specifically the Hough transform and its variants in addition to being robust to image transformations such as rotation, scaling and translation.

Letters: (2D)2 FLD: An efficient approach for appearance based object recognition

In this paper, a new technique called 2-directional 2-dimensional Fishers Linear Discriminant analysis ((2D)^2 FLD) is proposed for object/face image representation and recognition. We first argue that the standard 2D-FLD method works in the row direction of images and subsequently we propose an alternate 2D-FLD which works in the column direction of images. To straighten out the problem of massive memory requirements of the 2D-FLD method and as well the alternate 2D-FLD method, we introduce (2D)^2 FLD method. The introduced (2D)^2 FLD method has the advantage of higher recognition rate, lesser memory requirements and better computing performance than the standard PCA/2D-PCA/2D-FLD method, and the same has been revealed through extensive experimentations conducted on COIL-20 dataset and AT&T face dataset.

Rapid and brief communication: Visual learning and recognition of 3D objects using two-dimensional principal component analysis: A robust and an efficient approach

Inspired by the conviction that the successful model employed for face recognition [M. Turk, A. Pentland, Eigenfaces for recognition, J. Cogn. Neurosci. 3(1) (1991) 71-86] should be extendable for object recognition [H. Murase, S.K. Nayar, Visual learning and recognition of 3-D objects from appearance, International J. Comput. Vis. 14(1) (1995) 5-24], in this paper, a new technique called two-dimensional principal component analysis (2D-PCA) [J. Yang et al., Two-dimensional PCA: a new approach to appearance based face representation and recognition, IEEE Trans. Patt. Anal. Mach. Intell. 26(1) (2004) 131-137] is explored for 3D object representation and recognition. 2D-PCA is based on 2D image matrices rather than 1D vectors so that the image matrix need not be transformed into a vector prior to feature extraction. Image covariance matrix is directly computed using the original image matrices, and its eigenvectors are derived for feature extraction. The experimental results indicate that the 2D-PCA is computationally more efficient than conventional PCA (1D-PCA) [H. Murase, S.K. Nayar, Visual learning and recognition of 3-D objects from appearance, International J. Comput. Vis. 14(1) (1995) 5-24]. It is also revealed through experimentation that the proposed method is more robust to noise and occlusion.

Object recognition through the principal component analysis of spatial relationship amongst lines

This paper introduces a novel scheme which works on symbolizing every line in an object image for object recognition. Symbolizing is accomplished in terms of angles of intersection with regard to a line under consideration. Spatial relationship existing among the symbolized lines is represented using the notion of Triangular Spatial Relationship (TSR). A set of quadruples which preserves the TSR is subjected to principal component analysis to obtain the principal component vectors. These vectors are then stored in the knowledgebase for the purpose of recognition. Experimental results demonstrate that the proposed approach is efficient, invariant to linear transformations and capable of learning. To substantiate the success of the proposed model, a comparative study is performed with Murase and Nayar approach.

Eigen transformation based edge detector for gray images

In this paper, a simple and a robust algorithm to detect edges in a gray image is proposed. The statistical property of the small eigenvalue of the covariance matrix of a set of connected pixels over a small region of support is explored for the purpose of edge detection. The gray image is scanned from the top left corner to the bottom right corner with a moving mask of size k xk, for some integer k. At every stage, the small eigenvalue of the covariance matrix of the connected pixels that are having approximately same intensity as that of the center pixel of the mask is computed. This small eigenvalue is used to decide if a pixel is a potential edge pixel based on a pre-defined threshold value. The set of all identified potential edge pixels are then subjected to a pruning process where true edge pixels are selected. Experiments have been conducted on benchmark gray images to establish the performance of the proposed model. Comparative analysis with the Canny edge detector [1] and Sun et al. [15] is made to demonstrate the implementation simplicity and suitability of the proposed method in vision applications.

Appearance based object recognition using two-dimensional optimal feature transform

This paper proposes a new method of feature extraction called two-dimensional optimal feature transform (2D-OFT) useful for appearance based object recognition. The 2D-OFT method provides a better discrimination power between classes by maximizing the distance between class centers and minimizing the intra-class distance. We first argue that the proposed 2D-OFT method works in the row direction of images and subsequently we propose an alternate 2D-OFT which works in the column direction of images. To straighten out the problem of massive memory requirements of the 2D-OFT method and as well the alternate 2D-OFT method, we introduce bi-projection 2D-OFT. The introduced bi-projection 2D-OFT method has the advantage of higher recognition rate, lesser memory requirements and better computing performance than the standard PCA/2D-FLD/Generalized 2D-PCA method, and the same has been revealed through extensive experimentation conducted on COIL-20 dataset and AT&T face dataset