Conant Pablos, Santiago Enrique2018-12-172018-12-172018Casar, Alejandro (2018). Intelligent Input Dimensionality Reduction for Thermographic Computer-Assisted Breast Cancer Detection. ITESM, Monterrey, Nuevo León, Méxicohttp://hdl.handle.net/11285/632514While traditional breast cancer thermography consists of a bidimensional image of the breast area, this study explores the possibility of reducing the size of the thermal input required by thermography to the temperature at only a few points in the breasts. That kind of information could be retrieved by locating temperature sensors around the breast, which could mean a wearable breast cancer detection implementation. In order to do this, the document first briefly summarizes and explains the methodologies used by different authors in order to implement computer-assisted thermography. Then it proposes a topological framework used to translate thermal images into a subset of points that represent what a wearable device that meets certain physical constraints would look like. This is implemented by using thermal images from 167 patients, extracting the temperature by software at 120 proposed sensor locations, 60 on each breast, and storing it in the new proposed representation. This new framework is then used to compare the methodologies used by authors in previous works, but now all under the reduced-data representation, and all using the same patient database. After finding out which kind of data representation leads to better cancer detection, the work then proceeds to use greedy search algorithms to find the subset of sensors that maximizes the predictive power of a cancer detection classifier. After the experiments, an accuracy of 88.9% was achieved using only a subset of the temperatures at only 16 locations on each breast, and an accuracy of 90.1% was achieved when combined with some statistical and spatial features computed from the temperatures. In order to evaluate the validity of these promising results, a noise and robustness test was also carried out by introducing noise into the data and studying the effects that followed. The results of this robustness test were quite relevant as well since they showed that a backward elimination search strategy can be used to remove the most noisy-prone sensors, and a forward selection strategy could potentially be used to find an optimal sensor configuration that maximizes the predictive power of the classifier. Overall, the results from this work show that thermography with reduced dimensionality achieves performances that compete with full-image thermography, and even with mammography, with the additional advantage of being implementable through a wearable device, thus not requiring the patient to visit a hospital to perform the screening. Currently the only other portable massive self-examination method is the self breast-exploration, which yields an accuracy of around 47%, which is pretty close to being a random guess. Compared to this performance, the proposed method proves to be vastly superior. It could potentially replace self breast examination as the main first-hand portable, non-invasive and non-radiative breast cancer screening tool and provide a major breakthrough in the battle against cancer by detecting the disease in the earlier stages with a greater accuracy and consistency.Abstract ix List of Figures xii List of Tables xiv 1 Introduction 1 1.1 Brief Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Problem Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2.1 Breast Exploration . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 1.2.2 Mammography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 1.2.3 Thermography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.2.4 Room for improvement . . . . . . . . . . . . . . . . . . . . . . . . . 6 1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 1.4 Hypothesis, Objectives and Contributions . . . . . . . . . . . . . . . . . . . 9 1.4.1 Specific goals and questions . . . . . . . . . . . . . . . . . . . . . . 9 1.4.2 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 1.5 Document Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2 Theoretical Framework 13 2.1 Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.1 Supervised Learning . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.1.2 Feature Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 2.1.3 Cross Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 2.2 Performance Metrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2.1 Basic measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.2.2 Compound Measures . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.3 Previous Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.3.1 Evolution of computer assisted thermography . . . . . . . . . . . . . 19 2.3.2 Thermographic based feature extraction models . . . . . . . . . . . . 20 2.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 3 Solution Model and Methodology 27 3.1 Solution Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 3.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 4 Data Extraction and Transformation 31 4.1 Thermal Image Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.2 Discrete representation simulation . . . . . . . . . . . . . . . . . . . . . . . 32 4.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 5 Patient Representation Databases 37 5.1 Discrete patient representation adaptations . . . . . . . . . . . . . . . . . . . 37 5.2 Best Patient Representations . . . . . . . . . . . . . . . . . . . . . . . . . . 42 5.2.1 All vs All analysis - Friedman test . . . . . . . . . . . . . . . . . . . 43 5.2.2 One vs All analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 5.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 6 Optimal sensor and feature set configurations 53 6.1 Branched topology optimization . . . . . . . . . . . . . . . . . . . . . . . . 53 6.1.1 Individual branch impact . . . . . . . . . . . . . . . . . . . . . . . . 54 6.1.2 Branch combinations . . . . . . . . . . . . . . . . . . . . . . . . . . 56 6.2 Ring topology optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.2.1 Individual ring impact . . . . . . . . . . . . . . . . . . . . . . . . . 59 6.2.2 Ring combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 6.3 Sensor and Feature Selection . . . . . . . . . . . . . . . . . . . . . . . . . . 61 6.3.1 Sensor Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 6.3.2 Feature selection and combination . . . . . . . . . . . . . . . . . . . 62 6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 7 Robustness - Noise Resistance Check 65 7.0.1 Branch noise resistance . . . . . . . . . . . . . . . . . . . . . . . . . 66 7.0.2 Ring noise resistance . . . . . . . . . . . . . . . . . . . . . . . . . . 66 7.0.3 Sensor combination noise resistance . . . . . . . . . . . . . . . . . . 68 7.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 8 Conclusions 71 9 Appendixes 75 9.1 Appendix A: Visual Lab data acquisition procedure . . . . . . . . . . . . . . 75 9.2 Appendix B: Hybrid analysis with time series . . . . . . . . . . . . . . . . . 77 9.3 Appendix C: Database classification results . . . . . . . . . . . . . . . . . . 78 Bibliography 84104TextoengOpen Accesshttp://creativecommons.org/licenses/by-nc-nd/3.0/us/7 INGENIERÍA Y TECNOLOGÍATesis de maestríainput reductioncomputer-assistedbreast cancerdetectionthermographyscreeningIngeniería y Ciencias Aplicadas / Engineering & Applied Sciences