Feature Of The Week 11/26/2016: Tsinghua University Demonstrates Optical Computing Technique and Achieves High Speed OCT of 10 Mega-A-Scans/Second.
In the past few decades, OCT has been applied in a wide range of applications including clinical and material research areas. For some biomedical applications such as surgical guidance, real-time volumetric OCT (4D-OCT) imaging is in high demand. However, for high definition volumetric imaging in real time, i.e., 1000 pixels along 3 spatial dimensions and 30 volumetric imaging per second, the data flow will be 1000×1000×1000×30=30GVoxel/s. To process these massive data is a big challenge to the performance of current GPUs and CPUs. Professor Ping Xue, Postdoc Xiao Zhang and their colleagues from Tsinghua University in China, have made important progress to solve the problem. They have found a new way to implement the optical processing of optical spectrum and therefore optically solve the big challenge of massive data processing by GPUs and CPUs. In their paper, they reported a novel optical computing technique for high speed real-time OCT, as called OC-OCT or OC2T. Different from traditional FD-OCT, the interpolation in the wavenumber domain and FFT by GPU or CPU are not needed any more in OC2T. Instead of software algorithm and GPUs computing, they employed an all-optical real-time Fourier transformation system to process the A-scan data before the OCT light signal was converted to electrical signal by the photodetector. Using this optical computing technique, the A-scan rate and A-scan processing speed of OC2T have reached 10 mega-A-scans/second in their experiment, which is the fastest real-time sampling and processing speed to date. Furthermore, OC2T based on the all-fiber configuration is much more stable and easier to be used. The OC2T is able to be fast enough to achieve the video rate 4D-OCT imaging at a large volume size and therefore further enable the real-time virtual reality, which is of great value in surgical guidance. It is believed that this new technology will initiate new development of ultra-fast OCT technology.
For more information see recent Article. Courtesy Ping Xue from Tsinghua University.