Pattern Recognition in Information Systems 

About: Pattern Recognition in Information Systems is an academic conference. The conference publishes majorly in the area(s): Pattern recognition (psychology) & Computer science. Over the lifetime, 185 publications have been published by the conference receiving 984 citations.

Evaluating Pattern Recognition Techniques in Intrusion Detection Systems.

Abstract: Pattern recognition is the discipline studying the design and operation of systems capable to recognize patterns with specific properties in data sources. Intrusion detection, on the other hand, is in charge of identifying anomalous activities by analyzing a data source, be it the logs of an operating system or in the network traffic. It is easy to find similarities between such research fields, and it is straightforward to think of a way to combine them. As to the descriptions above, we can imagine an Intrusion Detection System (IDS) using techniques proper of the pattern recognition field in order to discover an attack pattern within the network traffic. What we propose in this work is such a system, which exploits the results of research in the field of data mining, in order to discover potential attacks. The paper also presents some experimental results dealing with performance of our system in a real-world operational scenario.

Incremental Non-negative Matrix Factorization for Dynamic Background Modelling

Abstract: A buoy flotation system, wherein a container of a compressed fluid is utilized for expansion of a float, includes a cutter for severing and removal of a portion of the container for releasing the compressed fluid. The cutter includes a piston having a recess in a side wall thereof for mating with the container, the piston being contained within a housing, the piston and the housing having cutting edges. An explosive charge drives the piston transversely of the container whereupon the cutting edges sever and remove a portion of the container for evacuation of the compressed fluid therefrom.

A Spatial Method for Watermarking of Fingerprint Images

Abstract: This paper extends the watermarking method introduced in [1] in order to embed watermark data into fingerprint images, without corrupting their features. Two methods are proposed. The first method inserts watermark data after feature extraction thus prevents watermarking of regions used for fingerprint classification. The method utilizes an image adaptive strength adjustment technique which results in watermarks with low visibility. The second method introduces a feature adaptive watermarking technique for fingerprints, thus applicable before feature extraction. For both of the methods, decoding does not require original fingerprint image. Unlike most of the published spatial watermarking methods, the proposed methods provide high decoding accuracy for fingerprint images.

Reversing and Smoothing the Multinomial Naive Bayes Text Classifier.

Abstract: Lehrstuhl fu¨r Informatik VI, RWTH Aachen – University of TechnologyD-52056 Aachen (Germany)ney@informatik.rwth-aachen.deAbstract. The naive Bayes text classifier has long been a core tech-nique in information retrieval and, more recently, it has emerged as afocus of research itself in machine learning. This paper is concerned withthe naive Bayes text classifier in its multinomial model instantiation.This model and an “equivalent” reversed version proposed here are in-terpreted under the statistical framework of log-linear modelling. Thereversed version provides an alternative way for parameter estimation,which (in a broad sense) is actually the main issue considered. The pa-per is to a large extent devoted to the study of the effects of parametersmoothing and document length normalization. For the purpose of pa-rameter smoothing, we consider a standard smoothing method for textclassification and two alternative techniques that are often used in thecontext of statistical language modelling for speech recognition. Empiricalresults are provided comparing these techniques and the effect of lengthnormalization for both the multinomial model and its reversed version.