Computer Assisted Cytological Medical Image Analysis

点击翻译按钮获取中文摘要

The proposed research leverages advances in Pattern Recognition, Image Processing, Computational Geometry and Machine Learning to develop novel solutions to address problems in the area of Medical Imaging. Specifically, I am furthering my research on Computer Aided Medical Diagnosis for cytological imagery of breast cancer fine needle biopsies, and of thin blood smears for malaria detection. For these types of cytological images, my long term objective is to develop robust, accessible, rapid, and accurate Medical Diagnosis tools to provide an automatic independent second opinion, to help avoid errors/oversights, to Pathologists, Clinicians and Medical Specialists, or act as an expert system to aid the non-specialist medical practitioner. The process of computer-assisted medical diagnosis or malignancy grading can be described in two main stages: feature extraction from medical imagery and classification of the imagery with respect to specific medical issues. One objective of the proposed work is the determination of improved feature sets for specific types of cytological images. Many of the frequently used features used for medical image classification problems require accurate image segmentation prior to determining these features. Therefore, I am studying approaches to improve the accuracy and speed of the segmentation of cells in cytological images by extending my previous work ranging from the use of the Hough transform to using active contours models, to incorporate effective use of texture, per-pixel classifiers, and shape modeling. We are also extending this research to the accurate segmentation of new medical imaging technologies such as very large full slide virtual images, 3D images, and images created with extended focal imaging (EFI). Another objective of the proposed research is to improve the accuracy of the malignancy classification in cytological images of breast cancer fine needle biopsies. Specifically, we study the problems of malignancy diagnosis to determine whether the slide is malignant or benign, and of malignancy grading where we use the Bloom-Richardson malignancy grading. To accomplish this, with optimized feature sets, we will develop systems of classifiers that improve sensitivity rates, particularly with minimal rates of false negatives. Since the classifiers have to work in a clinical setting, we will adapt the classifiers to do automatic parameter tuning. The third objective of the proposed research is to improve the detection of malaria parasites in cytological images of blood smear slides. Building on our previous complete blood count framework, we will develop a computer-assisted malaria detection system to determine the presence of malaria and return accurate infection counts per RBC counts in thick blood smears; and to differentiate between the four species of the malarial parasite in thin blood smears. We will develop systems of classifiers with improved sensitivity rates. My work in Health Informatics contributes to the larger trend in medical imagery to develop computer assisted medical diagnosis tools to complement and facilitate the work of Clinicians and Pathologists. All developments are designed to use commonly available microscopy and computing resources. For example, in the rural areas of malaria stricken countries there are frequently no specialists to do microscopic screening of blood slides for malaria. Through the automatic analysis of these slides, the computer assisted expert systems proposed in my research would help fill the knowledge gap between a specialist with the necessary expertise and the practitioner actually examining cytological slides, thus leading to better patient care.