Laurence S. Wilson

Knowledge-based method for segmentation and analysis of lung boundaries in chest X-ray images.

We present a knowledge-based approach to segmentation and analysis of the lung boundaries in chest X-rays. Image edges are matched to an anatomical model of the lung boundary using parametric features. A modular system architecture was developed which incorporates the model, image processing routines, an inference engine and a blackboard. Edges associated with the lung boundary are automatically identified and abnormal features are reported. In preliminary testing on 14 images for a set of 18 detectable abnormalities, the system showed a sensitivity of 88% and a specificity of 95% when compared with assessment by an experienced radiologist.

Fast content-based image retrieval using quasi-Gabor filter and reduction of image feature dimension

This paper introduces a new approach to content based image retrieval by texture. There are three problems to solve: high computational time, handling high dimension data, and comparing images consistent with human perception. To decrease the computational, time, we present a new strategy to extract an image feature with high retrieval accuracy. We also propose how to reduce the image feature dimension using the reward-punishment algorithm, so any robust indexing methods can be used. By weighting the extracted image features, a system may perceive the image consistently with human perception.

Multiple classification ripple down rule and fuzzy sets in computer aided diagnosis

Abstract This paper presents a computer aided diagnosis system in chest radiograph to extract abnormalities and suggest the possible disease states using multiple classification ripple down rule (MCRDR) and fuzzy function. The system generates image features in numerical parameter values using low-level image processing and converts them to linguistic (symbolic) descriptions using fuzzy function. The system then applies MCRDR to diagnose possible diseases. We prove the knowledge of the system is maintained and easily increased by radiologist themselves. MCRDR uses the knowledge supplied by experts just in the context it was provided, that is, by following the sequence of evaluated rules.

Intelligent computer-aided diagnosis system for chest radiography

Abstract We present an intelligent computer-aided diagnosis (ICAD) system for chest radiography to merge radiologic findings or the extracted features into a diagnosis using Multiple Classification Ripple Down Rule (MCRDR) and fuzzy functions. The ICAD system is a semi-automatic system since the abnormalities in chest radiographs could be detected by a combination of the automatic system and radiologists. Based on these abnormalities, the ICAD system suggests the possible disease states.

Characterization of arterial plaque using intravascular ultrasound: In vitro and in vivo results

Intravascular ultrasound is commonly used to image arteries before and after percutaneous procedures to restore blood flow in vessels where plaque has reduced the size of the vessel lumen. There is growing evidence that the success of such procedures depends on the composition of the plaque. In this study, this problem has been addressed by applying tissue characterisation techniques to intravascular ultrasound data obtained from a set of in vitro specimens of known pathology using a radiofrequency data acquisition system interfaced to a commercial intravascular scanner. We found that the attenuation slope of plaque is significantly increased in areas of nonfibrous plaque. This increase in attenuation slope is sufficiently high to allow it to be parametrically imaged and combined with gray‐scale imaging as a color overlay. Comparison with histologic sections shows that nonfibrous plaque can be detected even when it is not visible on the gray‐scale image. In addition, specular reflections may be readily distinguished from calcifications. Limited in vivo studies are consistent with the hypothesis that there is a relationship between plaque dissection and areas (probably calcification) detected by this method.