A Scalable Architecture for Real-Time Synthetic-Focus Imaging

Abstract

A scalable architecture for forming real-time synthetic focus images is described, and the design of a 256-channel system using currently available technology is presented as an example implementation of the architecture. The parallelism of the system scales directly with the number of array elements, and the image computation rate for a given image size (in pixels) stays constant as the number of array elements is increased. The system leverages earlier work in the real-time generation of the required time-of-flight surfaces and allows either real-time image generation or iterative adaptive image generation from a single complete dataset.