Design And Application Of Some Artificial Intelligence Systems For Computer Aided Detection

Abstract

Artificial intelligence-based computer-aided detection has been a major field of newlineresearch in past few years with their applications in disease diagnosis, bioimage newlineclassification and several other emerging areas such as prediction of toxicity in food newlineitems as well as environmental toxicity, affecting human beings. Machine learning and newlinemost recent deep learning-based algorithms are used to train these computer-aided newlinedetection systems. Machine learning is a sub-field of an AI; whereas deep learning is newlinesubfield of machine learning. In comparison to earlier machine learning based newlinealgorithms, deep learning algorithms based on deeper artificial neural network newlinearchitectures are capable to learn from massive datasets using advanced graphical newlineprocessing units (GPUs). newlineIn this thesis, I will use machine learning/deep learning-based techniques and newlinedevelop AI-based computer-aided detection framework for their applications in newlinebiology, medicine and toxicology. All the applications presented in this thesis are bind newlinewith the common thread, i.e., AI (Artificial Intelligence). Images captured using newlinedifferent modalities are used as source of data for the development of these computeraided newlinedetection/classification/diagnosis systems. newlineThe whole thesis has been divided into three main sections based on the newlineapplication of computer-aided detection system; these application areas are: biology, newlinemedicine and toxicology. For these three different application areas, AI -based newlineframeworks have been proposed, models have been created and new algorithms have newlinebeen developed, to solve some of the most crucial problems in those areas. The newlineeffectiveness of these computer aided detection systems has been tested and compared newlinewith the some of the recent state-of-the art systems proposed by other researchers. newline