Feature Of The Week 07/22/2019: Tsinghua University Demonstrates A “Sparse Sampled” Smart CCD Camera for Compressed Sensing Optical Coherence Tomography.
Optical coherence tomography (OCT) is a promising imaging technique in medicine and biology, which allows for obtaining high-resolution, cross-sectional images of biological tissues. In the past decade, the imaging speed of OCT is consistently increasing, and currently OCT can acquire three-dimensional (3D) images in real time with imaging speed up to 10-MHz A-scan rate. With such a high imaging rate, OCT system has to acquire a large amount of data, which is too massive for data transmission and storage.
Compressed sensing (CS) technique is a powerful method to solve this problem. It exploits the sparsity property of the signal and can reconstruct accurate signal with just a few randomly under-sampled measurements, thus effectively reduce the amount of data in acquisition, transmission and storage. Most of the CS-OCT researches utilized conventional spectral domain (SD) OCT system. The spectral data were fully acquired by CCD camera and then under-sampled by software. Although the feasibility of CS algorithm could be demonstrated in this way, the acquired data were not really reduced and even most of them were wasted after acquisition. Therefore, to really reduce the acquisition data and achieve faster A-scan rate in CS-OCT, a “sparse sampled” CCD camera, which has randomly arranged pixels and hardwarely under-samples the spectrum, is of high demand. However, such a camera is not commercially available up to now and may be very complicated and time-consuming to develop and fabricate.
Professor Ping Xue, Dr. Wenchao Liao and their colleagues from Tsinghua University in China, have made important progress to solve the problem. They have invented a novel technique to develop a sparse-sampled CCD camera which is effective for CS-OCT system. The sparse-sampling is realized by adding a customer-designed mask into the optical path in front of the CCD camera of the SD-OCT system. The mask may be easily manufactured with the current nanotechnology and is highly cost effective. They set up a 4f lens system and the mask is precisely imaged at the plane of CCD pixels. Therefore the mask can randomly block the left half or the right half of each pixel to realize random under-sampling, achieving the resolution of CCD with pixels doubled. The amount of acquired spectral data is finally reduced and no acquisition data is wasted, resulting in higher A-scan rate due to less acquisition data. Furthermore, the signal-to-noise ratio (SNR) and the fall-off performance of the OCT system are greatly enhanced owing to the CS technique. It is believed that this novel technique of fabricating sparse-sampled camera will be very useful for compressed sensing related researches and applications.
For more information see recent Article. Courtesy Ping Xue from Tsinghua University.