Search

There are various methods in practice to analyze deception by the law enforcing departments. Analyzing handwritten document is still considered to be one of the reliable methods of detecting deception. Handwritten documents are usual and reliable source of communication even in today's scenario. The handwritten text is considered to reflect human's body mind coordination. Researchers state that human behaviour can be analyzed by studying a person's handwriting. The paper consists of studies and implementation of online Handwritten Behaviour Analysis in proxmox-server virtualization Environment. C# language was used to implement the Behaviour Analysis part and python is used to validate the classification accuracy on the collected data. We have extended the previous work to client server mode to highlight the part of online Behaviour analysis to identify the uniqueness in characteristics of any person to differentiate deceptive and true statements. This paper will report the development and deployment of an online virtual server based handwriting behaviour analysis system using SVM based supervised Machine learning algorithm for classification.

Handwritten text is potentially the most powerful and conventional means of personal authentication in Human Computer Interaction, with applications to be found in document analysis, deception detection, banking and many other areas. Handwriting is a complex perceptual motor task generating linguistic information. Characters reflect shape distinction needed to perceive different phonetic information of words. In this paper, we have tried to emphasize the role of perception and cognition in identifying unique characteristics of handwriting of any person to screen out deceptive and true statements as a computational model in the areas of Pattern Recognition and Human Computer Interaction. The paper reports the prototype development of a decision support system based on handwriting behavior analysis.

Selection of relevant features for classification from a high dimensional data set by keeping their class discriminatory information intact is a classical problem in Machine Learning. The classification power of the features can be measured from the point of view of redundant information and correlations among them. Choosing minimal set of features optimizes time, space complexity related cost and simplifies the classifier design, resulting in better classification accuracy. In this paper, tomato (Solanum Lycopersicum L) leaves and fruiting habits were chosen with a futuristic goal to build a prototype model of leaf & fruit classification. By applying digital image processing techniques, tomato leaf and fruit images were pre-processed and morphological shape based features were computed. Next, supervised filter and wrapper based feature selection techniques were adopted to choose the optimal feature set leading to small within-class variance and large among-class distance which may be of utter importance in building the model for recognition system of the tomato leaf and fruiting habit genre.

The issue of comparing classification algorithms on a data set to find optimal classifier has always been a demanding issue in Machine Learning studies. The goal of this work was to compare the performance of different classifiers tested on spatio-temporal features used to distinguish between true and distorted handwriting behaviour to detect deception using machine learning experiments. Dynamic handwriting features depict a writer’s individuality and neuro-psychic condition. Handwritten features share inter-class characteristics while a few of them are very unique to an individual writer. However, extraction of discriminating features & selecting them in decision making process with respect to an individual is really crucial and complex thing, which helps the construction of any classifier, aiming to build the decision support system. Spatio-temporal attributes carry significant discriminatory information in deception detection. In this work, a set of total twelve features (spatial and temporal with pressure measures) features were used from online handwritten data. Five different classifiers (Naive Bayes, Logistic Regression, Multi-Layer Perceptron, Random Forest and Support Vector Machine) were tested on these feature set with three separate factual descriptions: Person, Action & Event, in both true and distorted mode to develop a decision support system based on the outcome of the experiment on handwriting behaviour. Both Support Vector Machine and Logistic Regression technique have shown good performance in distinguishing between true and distorted handwritten contents.

Image Enhancer is an open source, portable graphic editor developed for Windows platform. It is equipped with an enriched set of digital imaging filters with advanced computer vision techniques embedded within, like Interest Point Detection (Susan Corner Detector), Linear Edge Detection (Simple, Sobel, Canny), Histogram Equalization, Dithering (Bayer, Burkes, Sierra, Jarvis Judis Ninke), Transforming to Polar images and vice versa etc. Image Enhancer was released under GNU Lesser General Public License (LGPL) and the software was made available from the Microsoft’s open source project hosting repository Codeplex (http://imageenhancer.codeplex.com). Image Enhancer was tested and hosted by several popular software archives like SoftPedia, CNET, Freeware Files, ZDNet, Soft Tango and others. A stable Release Candidate (RC) version has been made available in which some major modifications were done which were not present in the earlier Beta version. The download link for the Image Enhancer (both Release Candidate & Beta Version) from CodePlex repository is (http://imageenhancer.codeplex.com/releases).