Dissertation > Excellent graduate degree dissertation topics show

Multistage Medical Image Retrieval System Based on Multi-features

Author: WangWeiWei
Tutor: LuHongBing；LiaoZuoMei；ZhangGuoPeng
School: Fourth Military Medical University
Course: Applied Computer Technology
Keywords: Content-based medical image retrieval Multi-feature Multi-level search Region of interest Three-dimensional images
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 55
Quote: 0
Read: Download Dissertation

Abstract


With the development of medical imaging technology and the popularity of the hospital information network, the available clinical, teaching and research use of medical images is the rapid expansion of content-based medical image retrieval (Content-based Medical Image Retrieval CBMIR) system for its effective become a hot research field of medical image management and retrieval of large medical image data. In recent years, researchers at home and abroad to carry out a large number CBMIR achieved certain results, but there is still insufficient, such as portability, to retrieve a single model, there are differences with human subjective judgment. How to expand application range of CBMIR, to meet the demand for multi-user multi-level search, retrieve images provide a convenient and accurate means for physicians and researchers, is the key to the development of medical image retrieval technology. The key technology of CBMIR research, feature extraction methods based on the characteristics of different types of medical images, detailed comparative analysis. Build a medical image database based on the design of a multi-level multi-feature medical image retrieval system. The main research work are as follows: ① study medical image text automatic extraction of information technology. Through the study of digital imaging and communication standard (Digital Imaging and Communications in Medicine, DICOM) of patients, serial automatic extraction and entry and indexing the combination used in the text and content of the text messages DICOM files search, to narrow your search. ② study multi-feature fusion technology. Visual content in order to more fully describe the image, according to the characteristics of the image underlying characteristics, based on the principle of complementarity, the texture feature and shape feature a combination of global features combined with the regional characteristics. Different combinations of features designed for two-dimensional global image, two-dimensional interested regional and three-dimensional region of interest. (3) study multi-level search mode. Retrieval needs of various users, the design of a three retrieval modes: a retrieval for text-based retrieval, image by the DICOM text information matching pre-screening; two retrieval based on global image retrieval by The level overall visual image can be retrieved for the subsequent ROI retrieval narrow Look in. Three retrieval ROI-based retrieval, is used to retrieve the doctors are more concerned about similar lesion area. ④ study the three-dimensional medical image retrieval technology. Data for 3D colon polyps, a three-dimensional gray-level gradient symbiotic matrix, three three-dimensional characteristics of the Shape Index (Shape Index, SI), and curvature (the Curvedness, CV,), then the use of these features a three-dimensional ROI divided described, establish index, to provide users with a way to reflect lesions in a more complete three-dimensional retrieval means. Finally, colon CT images and bladder MR images, by multiple retrieval experiments the following conclusions: the advantages of multi-feature fusion technology to effectively take advantage of the different characteristics, expression of the visual content of the image, the search is better than retrieval results using a single feature; multistage retrieval mode in accordance with the different needs of the user gradually narrow your search to find the desired image retrieval accuracy rate is generally higher than the single-stage retrieval accuracy; The 3D ROI retrieval due to the use of the specific characteristics of retrieval better, more clinical diagnostic significance compared with the two-dimensional image retrieval. Experimental results show that the multi-level multi-feature medical image retrieval system can effectively text messages and visual content, global image and ROI, two-dimensional images and three-dimensional images combined to improve the accuracy of medical image retrieval, feasible and practical sex.

Related Dissertations

  1. Airborne LiDAR Filtering by Fusing Multi-features,TN958.98
  2. Computerized 3D Reconstruction of Cat Optic Nerve Using Serial Sections,R338
  3. 3D Face Recognition Using Ifnormation Fusion,TP391.41
  4. Machine Vision Research on Technology of LED Die Sorting,TP391.41
  5. Digital Watermarking Algorithm of Color Image Based on DWT,TP391.41
  6. Research and Implementation of Image Retrieval Based on Color and Texture Feature,TP391.41
  7. Cerebral Image Processing Based on the Region of Interest and Its Application,TP391.41
  8. Design of High Frame Rate CMOS Camera with the Function of Windowing Readout,TP391.41
  9. The Video Quality Assessment Technique Based on the HVS Characteristic,TP391.41
  10. Research on Video Compression Algorithm for Wireless Multimedia Sensor Network,TP212.9
  11. Diffusion Tensor Imaging Study of Idiopathic Epilepsy at 3.0T High Field Magnetic Resonance Scaner,R445.2
  12. Imaging Study about Distribution Sites of Lesion Areas of Pulmonary Fibrosis Induced by Paraquat,R595.4
  13. Research and Application of Content-Based Image Database Retrieval Technology,TP391.41
  14. Research on 3D Model Retrieval System Based on Multi-feature and Relevance Feedback,TP391.3
  15. Researches on Content-based Multi-View Video Coding,TN919.81
  16. Dynamic SPECT Imaging to Measure Liver Function Reserve Clinical Application and Research,R735.7
  17. Control Study of Dynamic Enhancement CT and 18F-FDG PET-CT About the Pathology Types of Lung Cancer,R734.2
  18. Research on Image Segmentation Based on Dynamic Fusion of Multi-feature,TP391.41
  19. Extract visual keyword vocabulary tree - based image retrieval and research,TP391.41
  20. Using wavelet texture and shape features integrated image retrieval system and its implementation,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