CUSUM-Based Algorithm for Segmentation of Noisy Images with an Application to Optical Coherence Refractometry

• Activity Number: 514 - Recent Advances in Imaging Statistics: Bayesian Methods and Beyond
• Type: Contributed
• Date/Time: Wednesday, August 2, 2017 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): John Lu* and Jeeseong Hwang and Hong Chen
• Companies: NIST and NIST and Baruch College, the City University of New York
• Keywords: Image segmentation ; Change detection ; Optical Coherence Tomography ; Refractive Index ; Nonlinear Regression ; Measurement Errors

Abstract

Refractive index is a key optical parameter of materials such as glasses, and is fundamental to optical design and for the understanding of light propagation. In collaboration with NIST physicist Jeeseong Hwang, data are collected using optical coherence tomography (OCT) on some reference material of known shape, we want to test the measurement capability of fixed thickness and refractive index using OCT and known physical laws (aka a nonlinear regression model of measurement data). To acquire the measurement data, we need to develop some high-throughput segmentation method for noisy images, and we investigate a CUSUM-based algorithm for change detection for this purpose, and this paper reports our progress in this problem. We will report on data analysis using image analysis packages developed in R and implement the developed algorithm for image segmentation that can be applied to high-throughput image data collection.