Biomedical Engineering and Bioengineering Commons^™

Deep Learning For Digitized Histology Image Analysis, Sudhir Sornapudi

Doctoral Dissertations

“Cervical cancer is the fourth most frequent cancer that affects women worldwide. Assessment of cervical intraepithelial neoplasia (CIN) through histopathology remains as the standard for absolute determination of cancer. The examination of tissue samples under a microscope requires considerable time and effort from expert pathologists. There is a need to design an automated tool to assist pathologists for digitized histology slide analysis. Pre-cervical cancer is generally determined by examining the CIN which is the growth of atypical cells from the basement membrane (bottom) to the top of the epithelium. It has four grades, including: Normal, CIN1, CIN2, and CIN3. In …

Design Of X-Ray Source For Real-Time Computed Tomography, Wesley William Tucker

Doctoral Dissertations

"The reduction of motion blur in computed tomography (CT) drives the current research for multisource CT. Due to their compact nature, the current multisource systems utilize stationary angled anodes. Unfortunately, these configurations neither simplify the imaging geometry, nor satisfy the need for managing the high thermal loads demanded by real-time CT (30 acquisition frames per second). To add to the current field of knowledge, two x-ray tube concepts are presented in this dissertation. First, a simulation of transient thermal analysis was performed on a compact transmission-type x-ray tube anode operating in pulse-mode. A correlation was found between deposited beam power …

Applications Of Machine Learning In Nuclear Imaging And Radiation Detection, Shaikat Mahmood Galib

Doctoral Dissertations

"The main focus of this work is to use machine learning and data mining techniques to address some challenging problems that arise from nuclear data. Specifically, two problem areas are discussed: nuclear imaging and radiation detection. The techniques to approach these problems are primarily based on a variant of Artificial Neural Network (ANN) called Convolutional Neural Network (CNN), which is one of the most popular forms of 'deep learning' technique.

The first problem is about interpreting and analyzing 3D medical radiation images automatically. A method is developed to identify and quantify deformable image registration (DIR) errors from lung CT scans …

Design And Development Of A Compact X-Ray Tube For Stationary Ct Architecture, Ashish Vighnahar Avachat

Doctoral Dissertations

"Multisource architectures enable sweeping one or more x-ray beams across the imaging field-of-view faster than physically moving a single x-ray source and/or a detector. Hence, these architectures are attractive for the applications in which temporal resolution plays an important role, for example, cardiac computed tomography (CT) or real-time CT. One of the recent developments in multisource architectures for CT imaging is stationary CT architecture, whereby two separate stationary arrays -- one for x-ray sources and one for detectors -- are utilized to sweep one or more x-ray beams along the gantry and acquire 360 degree projections. To have a stationary …

The Viability Of Advantg Deterministic Method For Synthetic Radiography Generation, Andrew Albert Bingham

Masters Theses

"Time sensitive and high resolution image simulations are needed for synthetic radiography generation. The standard stochastic approach requires lengthy run times with poor statistics at higher resolutions. The investigation of the viability of a deterministic approach to synthetic radiography image generation was explored. The aim was to analyze a computational time decrease over the stochastic method. ADVANTG was compared to MCNP in multiple scenarios including a Benchtop CT prototype, to simulate high resolution radiography images. By using ADVANTG deterministic code to simulate radiography images the computational time was found to decrease over 10 times compared to the MCNP stochastic approach"--Abstract, …

Computer Aided Diagnosis Of Oral Cancer: Using Time-Step Ct Images, Jonathan T. Scott

Masters Theses

"In medical imaging it is a very common practice to use a technique known as Time-Step imaging in patients who might develop cancer. Time-Step imaging it a very powerful technique, however it can lead to unmanageable amounts of image data. Previously the only way to search all of this data was to manually look through all of the files. This had to be done by trained professionals who knew what to look for within the images and make a judgment about the patient based on the images. This paper discusses the development of an algorithm to have a computer search …

Computer Aided Detection Of Oral Lesions On Ct Images, Shaikat Mahmood Galib

Masters Theses

"Oral lesions are important findings on computed tomography images. They are difficult to detect on CT images because of low contrast, arbitrary orientation of objects, complicated topology and lack of clear lines indicating lesions. In this thesis, a fully automatic method to detect oral lesions from dental CT images is proposed to identify (1) Closed boundary lesions and (2) Bone deformation lesions. Two algorithms were developed to recognize these two types of lesions, which cover most of the lesion types that can be found on CT images. The results were validated using a dataset of 52 patients. Using non training …

Cervical Cancer Histology Image Feature Extraction And Classification, Peng Guo

Masters Theses

"Cervical cancer, the second most common cancer affecting women worldwide and the most common in developing countries can be cured if detected early and treated. Expert pathologists routinely visually examine histology slides for cervix tissue abnormality assessment. In previous research, an automated, localized, fusion-based approach was investigated for classifying squamous epithelium into Normal, CIN1, CIN2, and CIN3 grades of cervical intraepithelial neoplasia (CIN) based on image analysis of 62 digitized histology images obtained through the National Library of Medicine. In this research, CIN grade assessments from two pathologists are analyzed and are used to facilitate atypical cell concentration feature development …

Significance And Analysis Of Milia-Like Cysts In Dermoscopy Skin Lesion Images, Sneha K. Mahajan

Masters Theses

“Milia-like cysts (MLCs) are dermoscopic structures frequently observed in seborrheic keratoses(SKs), which are the most common type of skin lesions. Diverse appearances of these skin lesions make them difficult to differentiate from melanoma, a deadly type of skin cancer. Classified by size into two main groups, starry MLCs and cloudy MLCs, the presence of these structures in a skin lesion has been known to help differentiate benign lesions from melanoma. Though the presence of cloudy MLCs is not exclusively associated with SKs, they can be a useful tool to differentiate SKs from melanoma. This research study determines the statistical occurrence …

Automatic Detection Of Polypoid Skin Lesions: The Squash Sign, Saurabh G. Karnik

Masters Theses

“Earlier detection of malignant melanoma can save lives. Automatic in-vivo methods of melanoma detection including smart-phone applications are now available, but specificity is low, i.e. too many benign lesions are detected as possible melanomas. Polypoids, elevated and rounded lesions are one class of benign lesions. These raised lesions with piecewise circular borders, predominantly intradermal nevi and cutaneous polyps (IDNs), are always benign. Variegated coloring present in polypoid lesions can lead physicians to perform an unneeded biopsy. Contact dermatoscope plates with gel squash these lesions, leaving a “squash sign” marker of benignancy which can be automatically detected. We present a method …