BSc/MSc project ideas
Functional neuroimaging data analysis with machine learning
Functional neuroimaging data is openly available from multi-center studies for investigation of various cognitive and mental health issues. A promising stream of research in this domain is the utilization of machine learning algorithms to study brain connectivity & the affects due to cognitive processes or mental disorders. This project concerns the study of recent clustering and classification algorithms for mining functional neuroimaging data. You should have an interest (and/or skills) in machine learning, brain understanding, and scientific research to work in this area.
Deep machine learning
I am interested to explore deep learning (in particular, deep convolution neural network) for analysis of medical images. This study can be facilitated with deep learning libraries like Theano, Tensor Flow, matconvnet. You should have an interest (and/or skills) in artificial intelligence, machine learning and scientific research to work in this area.
Diabetic Retinopathy Screening with Smartphone
Diabetes is the key cause of retinal blindness for millions of people globally. Current detection of diabetic retinopathy and related eye diseases from retinal images using fundus camera is expensive and inconvenient since the imaging equipment is not portable and requires expert human observer for diagnosis. There is a need to study & develop imaging algorithms that can allow retinal image acquisition with lens attached to a smartphone and processing done either remotely or instantly on the smartphone. This can enable wide screening & early diagnosis of diabetic retinopathy, in particular in rural and distant areas in the developing world.
Software system development for cardiac optical mapping
This project concerns the development of a graphical user interface and/or parallel processing routines for processing optical cardiac mapping data which concerns electrophysiological activity data of the heart. You should have interest (and/or skills) in software development and scientific research to work in this area.
Biomarker selection for disease diagnosis
This project concerns the selection of discriminant biomarkers, from a bigger list, for robust disease diagnosis. It aims to explore machine learning approaches for feature selection and classification to identify a short list of biomarkers which are robust and discriminant enough to aid in the diagnosis.
In addition, I am interested to work in general in the following areas: medical image segmentation for quantitative analysis, medical image enhancement for improved analysis, and medical/health informatics.
The BSc/MSc students are welcome to discuss own project ideas related to above areas.