Task-based optimization and performance assessment in optical coherence imaging

Optimization of an optical coherence imaging (OCI) system on the basis of task performance is a challenging undertaking. We present a mathematical framework based on task performance that uses statistical decision theory for the optimization and assessment of such a system. Specifically, we apply the framework to a relatively simple OCI system combined with a specimen model for a detection task and a resolution task. We consider three theoretical Gaussian sources of coherence lengths of 2, 20, and 40 μm. For each of these coherence lengths we establish a benchmark performance that specifies the smallest change in index of refraction that can be detected by the system. We also quantify the dependence of the resolution performance on the specimen model being imaged.