Optical snake-based segmentation processor with a shadow-casting incoherent correlator

Abstract

What is believed to be the first incoherent snake-based optoelectronic processor that is able to segment an object in a real image is described. The process, based on active contours (snakes), consists of correlating adaptive binary references with the scene image. The proposed optical implementation of algorithms that are already operational numerically opens attractive possibilities for faster processing. Furthermore, this experiment has yielded a new, versatile application for optical processors.

Two approaches, coherent and incoherent, for the optical implementation of active contours

“…As demonstrated previously, our optical snake-based processor is also effective with a further overlapping white Gaussian noise 31 . The proposed multichannel implementation also inherits from the noise robustness of the SIRS model.…”

“…We already validated this architecture 31 to process optical segmentations. The major drawback of this architecture is the slight diffraction of the first plane that is largely compensated by three fundamental advantages.…”

“…31 In this paper, we present an enhanced, higher-speed version, with 4 channels. This configuration takes account of the parallel processing capabilities inherent to incoherent optical processors 33 .…”

Active contour segmentation by use of a multichannel incoherent optical correlator

“…The optical correlation is computed by an incoherent correlator based on the shadow-casting principle. 20,21 We already validated this architecture 17 to process optical segmentations. The purpose of this paper is to describe how our prototype has been improved by taking account of the parallel processing capabilities inherent in incoherent optical processors.…”

“…A monochannel prototype processor with a shadowcasting incoherent correlator was previously implemented by our group. 17 In this paper, we present an enhanced, higher-speed version with 4 channels.…”