Click to See ALL Projects in One Page: https://liangbright.wordpress.com/author/liangbright/
My research interests are: biomedical data analysis for computer aided diagnosis and intervention, machine learning, computer vision, bioimaging, biomechanics, algorithms for sensors and robotics.
Big-Data challenge and Data Science: A huge amount of image data is created every day from microscopes in biological laboratories and from imaging machines (CT, MRI, …) in hospitals. Traditional way of manual processing and analysis becomes almost inadequate.Computer algorithms and software applications will then help biologists and clinicians to obtain and analyze information from daily collected data, which may lead us to new discoveries of the fundamental principles in biology and improve healthcare services.
Biomedical image analysis is to design algorithms and implement them as software applications that can automatically extract information from images of biological structures from small cells to large organs, and relate the information to object status.
By providing automatic tools to biologists to analyze biological images, such as cell images from fluorescence microscopy, the fundamental mechanisms of many biological processes can be investigated by using a large amount of image data.
By providing automatic tools to clinicians to analyze medical images, such as cardiac CT images, valuable and quantitative information about the patients can be obtained in the process of the diagnosis and treatment.
Machine learning, computer vision and biomedical analysis techniques are heavily used and developed in the field of biomedical image analysis. Computer vision techniques (e.g. object detection/segmentation and tracking) will allow to obtain object information (e.g. shape and motion) automatically from images. Biomedical analysis will provide more valuable information about the object (e.g. such as stress, strain, flow field). Machine learning will allow to use prior knowledge to facilitate information extraction from images and relate the information to object status in order to make a diagnostic decision or preoperative surgical plan.
Algorithms for image analysis are based on many mathematical methods, such as linear systems, optimization, geometric modeling, stochastic process, information theory, machine learning, and so on. I also design new math-methods and extend existing math-methods whenever necessary. These algorithms need to be implemented as software applications, so we can use them. I use C++, Python and Matlab for programming. Also, a powerful computer is essential.
Algorithms are also needed to analyze signals from sensors, e.g., gas sensors, and/or guide robots to specific locations, e.g., locating a hazard odor source.