Informatics Publishing Limited

Neetesh Gupta ¹, Niket Bhargava ², Md. Ilyas Khan ¹, Shiv Kumar ³, Bhupendra Verma ⁴

Affiliations
1 Department of Information Technology, Technocrat Institute of Technology, Bhopal (M.P.), IN
2 Department of Computer Science, Bansal Institute of Technology, Bhopal (M.P.), IN
3 Department of Information Technology, Technocrat Institute of Technology-Bhopal (M.P.), IN
4 Department of Computer Science, Technocrat Institute of Technology, Bhopal (M.P.), IN

Abstract

For the purpose of content-based image retrieval (CBIR) An up-to-date comparison of state-of-the-art low-level color and texture feature extraction approach is presented in this paper. The CBIR problem is identified by us because there is a need to search the huge databases having images efficiently and effectively. in this paper we suggest a color and texture feature extraction algorithms. Special attention is given for CBIR is the similarity measurement using correlation coefficient with distinct distance matrices properties. A New approach for image retrieval technique is proposed to improve retrieval performance, and reduce the extraction search times. Matching is performed between the test image and the object image and quality of matching is measured in terms of grading.

Keywords

Similarity, Measurement, CBIR, Low Level Feature Extraction, Correlation Coefficient.

Full Text

     

Priyanka Jaiswal ¹, Niket Bhargava ², Rajesh Shukla ³, Manoj Shukla ⁴

Affiliations
1 Department of Computer Science & Engg., BIST, RGPV, Bhopal, IN
2 BIST, RGPV, Bhopal, IN
3 SIRT-e, RGPV, Bhopal, IN
4 SAMCET, RGPV, Bhopal, IN

Abstract

Collaborative filtering is a technique for reducing information overload and is achieved by predicting the applicability of items to users. In collaboration Filtering we use recommendation system because Recommender system applies knowledge discovery techniques to the problem of making personalized recommendation for information. Products or services during a live interaction. these system especially the k-nearest neighbor collaborative filtering based once, are achieving widespread success on the Web. In clustering algorithms, the applicability is predicted by the weighted sum model of ratings of k nearest items. This paper considers a new approach to user-item clustering in collaborative filtering algorithm using Markov Model for recommender systems. We use metrics like prediction strength, and after getting Matrix we apply Markov model for prediction which gives better prediction than weighted sum.

Keywords

Collaboration Filtering, K-Means, Markov Chain Model, Recommender Systems.

Full Text