Nitin Bhatia

Hough transform based fast skew detection and accurate skew correction methods

The Hough transform provides a robust technique for skew detection in document images, but suffers from high time complexity which becomes prohibitive for detecting skew in large documents. Analysis of time complexity on various stages of skew detection process is carried out in this paper. A complete skew detection and correction process is divided into three parts: a preprocessing stage using a simplified form of block adjacency graph (BAG), voting process using the Hough transform and de-skewing the image using rotation. Skew correction phase, which is hitherto a neglected area, is analysed for the quality of de-skewed images with respect to the type of rotation. Fast algorithms for all the three stages are presented and exhaustive analysis on time complexity is conducted. It is shown that the overall time taken for the whole process is less than one second even for very large documents. It is also observed that time taken in rotation is as significant as in skew detection which is reduced with the help of fast algorithms using integer operations. While the BAG algorithm is found to be effective for documents with Roman script, it does not provide satisfactory results for Indian scripts where headline is a part of a script.

Clinical decision support system for dental treatment

Abstract Background In this research, a decision making system, based on fuzzy inference mechanism as proposed by Mamdani, is presented. Literature suggests that there is a lack of consistency among dentists in choosing treatment plan(s). So, this research work aims to facilitate the dentist decide the treatment plan(s) of the broken tooth. Methods An expert system based on fuzzy logic has been designed to accept inaccurate and vague values of dental signs and symptoms associated with the broken tooth. We designed a knowledge base with 60 rules and used Mamdani inference algorithm to decide the possible one or more treatment(s) and suggest the same to the dentist. Results The results proposed by the system are compared with the dentists’ suggestions. The Chi-square test of homogeneity is conducted on 100 randomly generated sample cases with the help of three professional dentists. It is found that the results produced by the system are consistent with the treatment plan(s) proposed by the dentists. Chi-square value of the test is 3.843565 which is less than the critical value which is 12.592. Hence, we are unable to reject the null hypothesis that assumes the two populations are homogeneous with respect to treatments. Conclusions The accuracy of the proposed decision support system for the treatment of broken tooth enhances the confidence level of the dentists while making decision regarding the treatment plan(s). Simple and interactive GUI makes it easy to use.

Improved Hough transform for fast Iris detection

Objectives: an important part of digital image processing is recognition of patterns. Iris recognition is an important application of the same. As a result of the fact that CHT lacks speed due to its high time complexity, it is useful to recognize valid regions and make these regions the only regions to process. Extracting a valid region from the image not only reduces the image storage and the quantity of operations, it also improves the speed of the process. The objective of the work is to describe a method for enhancing the speed of extracting circle i.e. iris from the image without compromising the accuracy of the technique. A modified version of Hierarchical Hough transform is also proposed which further reduces the time taken to detect iris. An exhaustive analysis is conducted for a large number of iris images and comparison results are shown. Method: The HT space is limited by dividing it into nine equal parts and then the center most space is extracted as the valid region. Then CHT and HHT transformations are applied on the valid region. Results: About 88.77% of maximum improvement and 87.989 % of average improvement is recorded in time of execution of proposed CHT in comparison to existing CHT. Whereas, run time of HHT and proposed HHT when compared, results show that the maximum performance rate is 86.649% and average performance rate is 46.81%. Conclusion: in case of HT base detection methods, execution speed is directly proportional to the number of edge pixels in input image. Restricting the number of pixels provides the advantage of considerably simplified computations, and hence faster algorithms. The work documents that an aggregate comparison of proposed transformations with existing transformations has achieved a faster yet accurate method to recognize iris.

Election Results Prediction System based on Fuzzy Logic

ABSTRACT Election is very popular word in the universe. As per the definition, Election is a process to select suitable from a group of candidates. But the process and properties may vary for different sectors. It has different types. Like local election, legislative election, parliamentary election, presidential election, senate election or election in a small group / union. As election is very famous same way election prediction is again not a new keyword. It has same long life as election. But still there is a challengeable task to predict accurate result. The use of fuzzy logic in social science to evaluate the prediction is the core part of this paper. A toolbox from MATLAB software named fuzzy logic toolbox is used for this purpose. Some input arguments are being considered which scaled using linguistic variables to predict the chances of winning along with chances of losing of a candidate. Keywords an opinion sampling and through successive discussions with Fuzzy Logic, Mamdani, Election Prediction Evaluator, Membership Function, Linguistic Variable

Cross-Disciplinary Applications of Artificial Intelligence and Pattern Recognition: Advancing Technologies

The need for intelligent machines in areas such as medical diagnostics, biometric security systems, and image processing motivates researchers to develop and explore new techniques, algorithms, and applications in this evolving field.Cross-Disciplinary Applications of Artificial Intelligence and Pattern Recognition: Advancing Technologies provides a common platform for researchers to present theoretical and applied research findings for enhancing and developing intelligent systems. Through its discussions of advances in and applications of pattern recognition technologies and artificial intelligence, this reference highlights core concepts in biometric imagery, feature recognition, and other related fields, along with their applicability.

Fuzzy logic based decision support system for loan risk assessment

This paper is concerned with risks associated with the loan Risk analysis and related adaptation strategies have grown in importance and complexity during last years. The large amount of data and information that needs to be handled and integrated requires specific methodologies and tools. So selecting a new loan applicant is a important issue of a financial organizations or Banks. There are so many different factors which effect on the selection of applicant who apply for loan. So, its a problem for Banks and other private financers to choose right applicant so that minimum risk involved during loan process using fuzzy logic. We control this problem in quicker and cheapest way using knowledge and experience of experts which are involved in this process and having knowledge about this field. This paper provides risk analysis based on fuzzy inference system. And provide GUI based interface tool regarding different parameters of an applicant which effect loan risk assessment. The analysis of experimental results of different applicants checks the correctness and consistency of decision Support system for correct decision making.