Keeping image processing in the optical domain

Abstract

ABSTRACT Image distribution is getting more and more processor demanding, due to the increased quality requirements and the need for real time imaging. Therefore, improv ed techniques and processo rs are being designed. Another attractive approach is keeping image processing in the optical domain and extending, as much as it is possible, the functionalities performed to this domain. Optical processing is a reality in many research fields, namely optical communications. Results reported in literature show that it is possible to exploit optical components to accomplish image capture, optical transform processing, sampling and threshol ding. We propose a general architecture for all-optical image acquisition, processing and transmitting, and re view potential best fit for each functionality. Keywords: All-optical processing, optical transforms, data compression, image/video processing. 1. INTRODUCTION At present, digital video or image delivery is one central researc h topic, since the technology is now mature to provide a concrete prospective of progress in this field, and also the interest on video streaming is clearly growing, from both services provider and users. Optical fiber networks are already used for image/video distribution, but photonic technologies can be exploited to accomplish not just delivery, but processing functionalities as well. This approach allows keeping image in the optical domain, with the first relevant achievement of overcomi ng the limitations due to Optical-to-Electrical-to-Optical (OEO) stages. In this framework, the main aim of the CITO (Image Co mpression using Optical Transforms ) project is to evaluate the potential of optical technologies in meeting the high processing requirements of actual and emerging image and video formats, and in overcoming limitation imposed by electronic based architecture, such as latency and power consumption. Therefore, with this work we aim to present a general architecture, for all-optical image/video acquisition, processing and transmitting. Proposed system architectu re is based on the following main fu nctionalities: video capture, transform, quantization, coding and transmission. A review of some of the potential best solutions and technologies for the implementation of such functionalities is presented, and the main challenges are investigated. To the best of our knowledge, the potential of integrated optics in the field of all optical video/image processing system is still not well exploited, especially from the integration point of view. Thus, the subject deserves major investigations and motivates our objectives.