Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device

The Segmentaion Research on Cancer Cells in Microscopy Images of Thyroid Gland

Author LiuChunXiang
Tutor DuanJun
School Inner Mongolia University of Science and Technology
Course Applied Computer Technology
Keywords Image processing Image recognition Image understanding edgedetection fuzzy recognition
CLC TP391.41
Type Master's thesis
Year 2013
Downloads 12
Quotes 0
Download Dissertation

In recent years, pattern recognition techniques in medical diagnosis becomeincreasingly widespread,which become a hot research topic in the field of artificialintelligence.. With the diversification of people’s lives, as well as environmental factors.Cancer incidence rate and morbidity has increased and diversified year by year, whichmakes it more difficult for the clinical diagnosis of the disease. It is an important means fordiseases to be diagnosed by microscopic image processing technology.when the image of the study begins, pre-treated should be conveyed to remove theinterference, noise and so on.If image information is so weak so that can not be displayedproperly, the need to enhance or transform processing for effective analysis of human-machine. To sum up, the image processing including image coding, image enhancement,image compression, image restoration, image segmentation etc, which segmentation part isso crucial that has a direct impact on the understanding of the results of the recognition.Image recognition is subsequent period after image processing and classification inorder to determine the Category. By the way of selecting the features extracted above, themeasurement of certain parameters, classifies basing on the measurement results.Structural analysis and image description should be made on the whole image in order toget an explanation on the image information and by the deep understanding of thestructural relationship of the multi-object image we can get a better result of recognition.The image deepen our understanding of the relationship between the structure of the multi-object, to achieve better recognition effect. For image recognition structure, the input isimage (treated), the output is a class and analysisImage understanding is an umbrella term for processing and recognition which is theultimate goal, to describe the image, and ultimately determine what image is. On the basisof the image processing and recognition, Classified as syntactic analysis, describe andexplain the image. Image understanding, include image processing, image recognition, and structural analysis. For the understanding input is image and the output is the descriptionand interpretation of the image, which belongs to AI fieldIn this thesis, the cancer cells in the thyroid microscopic images is for the study, Byproposing an improved Canny operator edge detection, segmentation algorithm andcombined it with image processing and recognition process, simplify identifying. Themain work is as follows:1、Learn image processing, recognition, understanding of the relevant knowledge,and program the common image processing algorithms and get the good result.2、By embedding an morphological improved Anti-noise operator into traditionalCanny operator, proposing an improved Canny edge detection method which proves out tobe better than traditional Canny operator in experiment.3、By exploring theory of feature extraction and recognition process for furtheridentification process. For the complexity of thyroid cancer cell which is hard recognizedWe proposed a recognition method based on fuzzy theory which lay a foundation onthe further study.

Related Dissertations
More Dissertations