MODELS AND ALGORITHMS FOR A REAL-TIME HYBRID IMAGE-ENHANCEMENT METHODOLOGY

The paper presents models and algorithms for a hybrid method of image enhancement. The proposed framework involves processing of the digitiz ed image to transform the input signal from image sensors such that th e digitization of the transformed analog signal yields the enhanced im age. First, an analytical framework which yields robust algorithms for image enhancement (following our approach) is established. Unlike man y conventional methods, the approach allows the user to specify certai n statistical measures for desired gray values in the enhanced image. An appealing feature of the proposed method is its ease of implementat ion using off-the-shelf hardware such as Analog-to-Digital Converters. To support our models and analyses, we perform character recognition on a set of industrial parts. Copyright 1996 Pattern Recognition Socie ty. Published by Elsevier Science Ltd.