Real-time optical imaging acquisition and processing in Python: a practical guide using CAS

Real-time data acquisition and processing is an important step in the development of new approaches to optical imaging in research laboratories. Python is increasingly used for scientific computing and allows for the straightforward application of artificial intelligence models using popular frameworks such as PyTorch. However, achieving high-speed image capture and processing in real time is challenging and requires extensive development work, a particular problem for academic labs where research teams may lack specialist expertise in software development. This note provides guidelines for achieving high performance in Python for optical imaging applications and introduces an open-source framework “CAS” for rapid prototyping of imaging system software. CAS includes a hardware abstraction layer for cameras, a ready-made GUI, which can easily be customized, as well as support for using multiple CPU cores for parallelism. By providing an open-source and flexible Python-based solution, CAS can support research teams to more quickly develop real-time imaging systems.