A method is described to improve the performance of sensor fusion algorithms. Data sets available for training fusion algorithms are often smaller than desired, since the sensor suite used for data acquisition is always limited by the slowest, least reliable sensor. In addition, the fusion process expands the dimension of the data, which increases the requirement for training data. By using structural risk minimization, a technique of statistical learning theory, a classiÞer of optimal complexity can be obtained, leading to improved performance. A technique for jointly optimizing the local decision thresholds is also described for hard-decision fusion. The procedure is demonstrated for EMI, GPR and MWIR data acquired at the US Army mine lanes at Fort A.P. Hill, VA, Site 71A. It is shown that fusion of features, soft decisions, and hard decisions each yield improved performance with respect to the individual sensors. Fusion decreases the overall error rate (false alarms and missed detections) from roughly 20 % for the best single sensor to roughly 10 % for the best fused result.

This paper presents a generalized statistical texture analysis technique for characterizing and recognizing typical, diagnostically most important, vascular patterns relating to cervical lesions from colposcopic images. Recognizing the fact that the texture patterns related to cervical lesions are primarily due to the vascular structures, the technique first extracts the vascular structures from the original cervical images, followed by vectorizing the extracted vascular structures with piecewise connecting line segments. Statistical distributions of the line segments are then constructed. First and second order statistics derived from the distributions are used as texture measures for cervical lesion classification. Experimental study demonstrated that the developed algorithm is very promising in discriminating between cervical texture patterns indicative of different stages of cervical lesions. Classification of cervical lesions consisting of six different vascular patterns yielded a...

We propose a new kernel direct discriminant analysis (KDDA) algorithm in this paper. First, a recently advocated direct linear discriminant analysis (DLDA) algorithm is overviewed. Then the new KDDA algorithm is developed which can be considered as a kernel version of the DLDA algorithm. The design of the minimum distance classifier in the new kernel subspace is then discussed. The results of experiments on two well-known facial databases show the effectiveness of the proposed method in face recognition. The results of experiments also confirm that DLDA can be viewed as a special case of the proposed KDDA algorithm. 1.

Steganographic tools and techniques are becoming more potential and widespread. Illegal use of steganography poses serious challenges to the law enforcement agencies. Limited work has been carried out on supervised steganalysis using neural network as a classifier. We present a combined method of identifying the presence of covert information in a carrier image using fisher’s linear discriminant (FLD) function followed by the radial basis function (RBF). Experiments show promising results when compared to the existing supervised steganalysis methods, but arranging the retrieved information is still a challenging problem.

Steganalysis plays an important role in identifying unacceptable information transmitted through internet communication system. In the process of steganalysis many untoward incidents can be avoided. Many techniques have been proposed and new techniques are tried with different combinations to maximize the efficiency of retrieving hidden information. We have proposed a combination of polynomial vector with Fisher’s discriminant function using the information of bitplane and radial basis neural network (PVDRBF). Each set of pixel is preprocessed to obtain interpolated pixels using PDV. This is further trained by Fisher’s discriminant method that transforms once again into 2-dimensional vector. A processing of training the RBF is adopted to obtain set of final weights. During implementation, the final weights are used to classify the presence of hidden information. Key words: