Dissertation > Excellent graduate degree dissertation topics show

Gastric Adenocarcinoma Microscopic Image Segmentation Algorithm Research

Author: ZhangPuSheng
Tutor: LiuJianPing
School: National University of Defense Science and Technology
Course: Control Science and Engineering
Keywords: Medical Cell Image Segmentation Membership FCM algorithm Mathematical Morphology Iteration corrosion Database
CLC: TP391.41
Type: Master's thesis
Year: 2007
Downloads: 76
Quote: 4
Read: Download Dissertation

Abstract


Computer technology has been widely used in medical diagnosis, which uses pattern recognition technology to automatically recognize cancer cells microscopic images is one of the important applications in the medical field. In this paper, HE staining of gastric adenocarcinoma cell microscopic image as the main research object, build on the preliminary study stage gastric adenocarcinoma aided diagnosis system was perfect. 6 wherein the texture features extracted by analyzing the various features of cancer cells, the three color characteristics and the morphological characteristics, as the basis of classification aided diagnosis system of cells. Cells were stained with difficulties and complex images automatically identify is not easy to achieve, the study on the basis of many mature segmentation techniques adapt gastric adenocarcinoma cell division improved algorithm: First, accurate segmentation, extraction of characteristic parameters, calculate the nucleus , proposed an FCM algorithm based on the improvement of the membership function. Consider traditional FCM algorithm does not involve spatial relationships into membership settings to address the shortcomings of the traditional methods of image noise sensitive. Neighborhood pixels by introducing the characterization of the role of the a priori probability of the center pixel to re-determine the fuzzy membership value of the current pixel, the probability of the algorithm execution process based on fuzzy membership value be determined automatically. The experiments show that the new algorithm can get more than the traditional FCM algorithm reasonable classification results. Second, accurate segmentation of the overlapping cells appear in the image, the paper uses a separation algorithm based on the weight of the morphological separation region overlapping cells. Obtained by improving the method of setting of the morphological structure elements, the closer to the ideal shape of the structural elements, to solve the false limit corrosion pixel aggregation; obtain accurate seed points and the associated maximum separation zone, and ultimately get a reasonable estimate of the outline of a single cell. By the new algorithm is integrated into the original system platform to improve the original segmentation algorithm insufficient. The tests showed that the improved FCM algorithm to get the classification results more reasonable image of the cells, and the overlapping cell separation algorithm can be based on a reasonable outline estimated to provide the correct classification of the basis for the final identification of the cell. Finally, the re-design of the original platform database module, perfect aided diagnosis system functions constitute a large collection of data for teaching practice and further research.

Related Dissertations

  1. Design and Development of Application Programme for Fault Analyzer Based on Wince Platform,TP311.52
  2. Study on System of Electronic Documetnt Security Signature,TN918.2
  3. Research on Compression, Operation and Query Processing Methods of Massive Datasets,TP311.13
  4. Tongue Feature Extraction and Research of Fusion Classification,TP391.41
  5. Implementation of Data Compression, Operation and Query Processing System Based on BAP,TP311.13
  6. Design and Implementation of the weapons and equipment information management system,TP311.52
  7. Force Online Examination System Design and Implementation,TP311.52
  8. The Properties of Laser Speckle Based on the Mathematical Morphology,O29
  9. Xinjiang Oilfield Ground Construction Cost Indicators and Management Information System Study and Application,F284
  10. The Studying of University Students’ Community Management Based on GIS,G647
  11. A Study on Healthcare Product Marketing Based on Data Mining Technology,F426.72
  12. Research and Design of One Kind of Paper’s QCS That Based on Embedded System,TP368.1
  13. Development of the on-line Training and Examination System of Army,TP311.52
  14. Road extraction algorithm based on region segmentation of remote sensing image,TP751
  15. Motor assist system for motorized units,TP311.52
  16. Online examination system design of military theory and,TP311.52
  17. Technology of Blood Vessel Diameter Measurement Automatic Based on Digital Image Processing,R310
  18. Tibet border combat training information management system design and realization of research,TP311.52
  19. The Establishment of Chemical Toxicological Database for Wool Industry,TS131
  20. Structure of Wenzhou Agricultural Resource Management Information System Based on GIS,S126
  21. The Building and Application of Jinqu Basin Soil Series Database of Zhejiang Province,S159.2

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