Dissertation > Excellent graduate degree dissertation topics show

Research on Some Technologies of Local Feature Image Retrieval

Author: LiQiZhou
Tutor: ChenWenBing
School: Nanjing University of Information Engineering
Course: Applied Mathematics
Keywords: Image Retrieval Local features Mean Shift Contour curves Target tracking
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 92
Quote: 1
Read: Download Dissertation

Abstract


Content-based image retrieval (CBIR) using the image itself, with the help of the existing image processing techniques and to construct a new algorithm to identify the mechanism of the image features, and compare the characteristics based on each image can be retrieved. Currently a lot of automatically based on the content of image retrieval are global image information, in most cases, however, the user is more concerned is a region of the image having a certain semantic, in order to achieve this effect, some of the image retrieval system, the introduction of image segmentation and automatic region extraction technologies, not yet, however, there is a universal way, also do not have a judgment of segmentation quality standards, and thus split the results will inevitably cause with people's subjective understanding of the differences, but also can not be accurately extract relevant visual features of the region, but also reduces the effect of the search results. In order to solve the above problems, the paper mainly research oriented image retrieval based on local features. The study of specific methods are as follows: an integrated color and image contour curves retrieval method. The method first divided image and extract the contour of the object of interest in the image, then the affine transformation on the extracted outline and minimum processing with a complete information on the edge of the contour of the treated and having a geometric invariance; Secondly, The clustering of color information, to extract a histogram of the main cluster, the histogram extracted not only contains the color information of the master cluster also includes the spatial location information of the cluster. Finally, the weighted average metric retrieval and similarity of the objects to be retrieved using the search object and the color of the object is retrieved from the histogram and the contour curve distance deviation. Retrieval method based on the region of interest, the user needs to select the objects you want to retrieve their own picture as the search object. The algorithm Mean shift tracking ideas use the content-based image retrieval, but the classic Mean Shift tracking algorithm using color histogram to track the target, and did not take into account the scale of change and the spatial location of the target pixel response to these problems, this paper a fast adaptive algorithm to adjust the window width scale, the algorithm can be fast, accurate search retrieval images to change the size of the target window width scale, followed, in the the traditional histogram of spatial information, the information reflecting the spatial position of the target pixel to improve the tracking robustness, and finally, the use of the color distribution of entropy to measure the similarity of the two pictures, the method is more truly reflect the spatial structure of an object. The experiments show that the classic mean shift to do a series of improvements, can improve the tracking accuracy of the positioning accuracy and retrieval effectiveness. The development of a content-based image retrieval engine, the software implementation of a content-based image retrieval method. The software is based on the B / S architecture, the software according to the different characteristics of the image retrieval and retrieval needs, to achieve four different image retrieval method, which achieved the main cluster matching-based image retrieval method is the result of the research team of independent innovation, its main advantage is that the algorithm extracts a feature not only limited to a single feature extraction, but the combination of the multiple features of the image, thereby improving the accuracy of the image retrieval.

Related Dissertations

  1. The Maneuvering Target Tracking Research Based on VRPF,TN957.52
  2. Research on Multi-Sensor Embattling and Continious Tracking Technique Applying for Anti-Stealth and Anti-Interference,TN953
  3. Research on Facial Feature Extraction and Matching Algorithms for Image Retrieval,TP391.41
  4. The Research and Implemention of Image Retrieval Based on User Interested Feature,TP391.41
  5. Application of Q-Learning in the Content-Based Image Retrieval Technology,TP391.41
  6. Research and Implementation on Content-Based Clothing Image Retrieval,TP391.41
  7. The Discovery of User Concept Region Based on Multiple Instance Learning,TP391.41
  8. Research on Transductive Support Vector Machine and Its Application in Image Retrieval,TP391.41
  9. Research and Application of Diverse Density Learning Algorithm,TP181
  10. Tracking Cells in High Density Image Sequences Based on Mean Shift Algorithm Combined with Topological Constraint,Q25
  11. Algorithm Analysis about Traget Tracking in Wireless Sensor Network,TN929.5
  12. Research on Object Detection and Tracking Method in Active Vision System,TP391.41
  13. Research on Moving Object Retrieval for Video Surveillance,TP391.41
  14. The Study of Moving Object Detection and Tracking Algorithm Based on Image Information,TP391.41
  15. Research of Moving Object Detection and Tracking Technology,TP391.41
  16. The Research on Localization and Target Tracking in Wireless Sensor Network,TN929.5
  17. Research of Image Segmentation in Web Image Search Based on GPUs,TP391.41
  18. Video image sequence moving target acquisition and tracking,TP391.41
  19. Mean-Shift -based KLT and target tracking study,TP391.41
  20. Content-based image retrieval technology research large-scale digital,TP391.41
  21. Based on embedded image tracking system,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile