Dissertation > Excellent graduate degree dissertation topics show

Image Retrieval Based on Relevance Feedback

Author: HuYing
Tutor: DingMingYue
School: Huazhong University of Science and Technology
Course: Pattern Recognition and Intelligent Systems
Keywords: Statistical Learning Theory Machine Learning Content-based Image Retrieval Transitive SVM Memory relevance feedback
CLC: TP391.3
Type: Master's thesis
Year: 2004
Downloads: 302
Quote: 6
Read: Download Dissertation


Intenet rapid development of technology, through a text message to retrieve multimedia information retrieval has been unable to meet the advanced needs. Image processing technology, provides a new way image retrieval: content-based image retrieval. Content-based image retrieval technology in the general image search engine, trademarks match retrieval, Museum Art image management and retrieval, web filter design has broad application prospects. The research projects aimed at developing a set with a certain combination of practicality SVM method and add relevance feedback mechanism for content-based image retrieval systems, and building systems based on the expanded image retrieval relevance feedback algorithm. This paper discussed the theory and practical aspects of support vector machine based relevance feedback mechanism to realize the problem: In the traditional training algorithm analysis, to determine a more effective training algorithm - SMO; addressed using different kernel functions for SVM Classification of the same database performance difference big problem for the image database through experiments selected the appropriate kernel function and the corresponding parameters. Built on content-based retrieval techniques in-depth analysis of the soil on the basis of this paper, the basic characteristics of retrieval algorithms: Characterization is vector-based, composed of different spatial search algorithm independent of each other and the different characteristics of the joint search by the weighting; in Based on this, discussed the relevance feedback image search engines to build a system problem, and implemented an experimental system. This realization of the SVM-based relevance feedback retrieval algorithms in the analysis of this method in the web-based image retrieval system based on the lack of proposed and implemented based on support vector machine transitivity HSVM (Hand-on Support Vector Machine ) with memory relevance feedback image retrieval. In the human-computer interaction process, HSVM classifier not only for this feedback process user submitted labeled positive examples and negative examples for learning, but also on previous feedback process of positive and negative examples samples for study, and according to training classifier retrieval. Experimental results show that the method in the case of small sample set, you can still retrieve more relevant images, in the case of limited training samples with good generalization ability, and its retrieval performance is better than traditional relevance feedback methods - MARS methods and SVM-based relevance feedback methods. Finally, the paper discusses the problem of image retrieval system construction, and implementation of an experimental system - basestar_new.

Related Dissertations

  1. The Research and Implemention of Image Retrieval Based on User Interested Feature,TP391.41
  2. Application of Q-Learning in the Content-Based Image Retrieval Technology,TP391.41
  3. Research and Implementation on Content-Based Clothing Image Retrieval,TP391.41
  4. The Discovery of User Concept Region Based on Multiple Instance Learning,TP391.41
  5. The Research on English-Chinese Name Entity Translation,TP391.2
  6. Based on Data Distribution Characteristics of Text Classification,TP391.1
  7. Prediction of Binding Affinity of Human Transporter Associated with Antigen Processing,R392.1
  8. Research on Content-based Trademark Image Retrieval Technology,TP391.3
  9. Research on Relevance Feedback Techniques Based on Long-term Log Learning and RW-Soft SVM in CBIR,TP391.3
  10. Research on Several Machine Learning Methods and Their Applications in Video-Based Fingerprint Verification,TP391.41
  11. Research on Flat Features and Structural Information for Protein-Protein Interaction Extraction,TP181
  12. The Research and Realization of the Military Port Objects Classification Platform,TP751
  13. Research on Locally Features Aggregating and Indexing Algorithm in Large-scale Image Retrieval,TP391.3
  14. Jade Multi-Agent Based Image Retrieval System,TP391.3
  15. Content-based image retrieval technology research large-scale digital,TP391.41
  16. Image retrieval method and system for parallel computing,TP391.3
  17. Sketch-based image retrieval technology research and system implementation,TP391.41
  18. Online Learning and Its Implementation in Intelligent Traffic and Finantial Market,U495;F830
  19. Research on Algorithms to Reassign Labels to Regions,TP391.41
  20. Research on Method of Retrieval from Remote Sensing Image with Global and Local Features,TP751
  21. Research on Color Image Retrieval Based on Visual Attention,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Retrieval machine
© 2012 www.DissertationTopic.Net  Mobile