High Speed Data Processing for Optical Coherence Tomography
Departments: Electrical and Computer Engineering Advisor: Chao Zhou
Optical Coherence Tomography is a non invasive three dimensional imaging approach that offers high resolution (a few microns), high speed, and 1-2mm penetration depth in biological tissue. It is most widely used in ophthalmology, and finds applications in other medical fields as well as some niche areas such as art restoration. For many reasons including relatively low equipment costs, the technology is also widely used in academic research. Functionally, it is often compared to ultrasound, in that it detects signals reflected by features of a sample. A Fourier domain approach to OCT has become dominant due to advantages in imaging speed and signal to noise ratio. However, transforming frequency domain volumes of raw data to user viewable images is a computationally intensive multi-step process, and due to the technology’s high imaging speed, is usually the bottleneck in a user’s workflow.
A software application was developed to increase OCT data processing speed. Written in CUDA C++, the application performs the necessary computations on an nVidia Graphics Processing Unit (GPU), allowing for a significant improvement in processing speed over the current CPU based solution. While the speed improvement is notable, it can be realized on a low cost workstation and provides results identical to the slower solution.
About Andre Sukernik:
Andre Sukernik is a Lehigh University senior studying Electrical Engineering and minoring in Nanotechnology. He has worked under Dr. Chao Zhou since Dec 2013, participating in a few different projects. After graduation, he plans to pursue a master’s degree in electrical engineering.