SenseAI - AI-driven compressed image acquisition

点击翻译按钮获取中文摘要

From satellites to CCTV to CT-scans, sensing systems of all kinds are becoming ever more sophisticated, but in doing so they are generating more data than ever, and creating ever more challenges as a result for bandwidth, data analysis and storage.SenseAI is a suite of patented technologies from the University of Liverpool that was developed from the ground-up specially to overcome these drastic limitations. At the core are 'subsampling' methods which exploit redundancy within data, and allow entire images and spectra to be formed from only a subset of measurements, meaning that hi-fidelity images can be acquired, reconstructed, and analysed using only a fraction of the time or radiation exposure (low as 1%).It originated from the field of Electron Microscopy (where the University of Liverpool has a world-class team and facilities) and where these challenges severely restrict usage, and beam damage to samples is also a key problem. Through the ICURe Explore process, other future markets for the application were discovered, in particular medical imaging (X-Ray, CT, NMR, MRI and ultrasound) as well as for security such as CCTV and military such as airborne sensors and satellite imaging. Though this project, SenseAI seeks to develop enterprise-grade software, to speed up and democratise image processing within many fields, and unlock more possibilities in analysis, health screening and security