Pieter Vuylsteke

Trainable rule-based network for irradiation field recognition in Agfa's ADC system

The irradiation field is the diagnostic region of radiograph which has been exposed directly to x rays and has not been shielded from the source by x-ray opaque material. Such material may have been placed to shield vulnerable regions of the patient from unnecessary exposure, and also to partition the radiograph into sub-images, thereby permitting multiple exposures on the same plate. In this paper we describe an image analysis method to automatically detect and locate the irradiation field of a digital radiograph. The aim is to prepare the radiographic image for contrast-enhancement processing which is driven only by exposure data from the diagnostically useful part of the radiograph. The algorithm is broadly edge-based, and comprises four stages: (1) edge detection, culminating in a set of straight-line edge segments as non-iconic data structures; (2) clustering (where possible) these segments into longer lines in accordance with colinearity constraints and the like; (3) model matching by a trainable rule- based network to identify irradiation field boundaries; (4) masking, which acts upon the input image to blank out the regions not in the irradiation field. The network has been trained on a database of 2100 images, and has been tested in clinical use, delivering an accuracy for recognition of the irradiation field of better than 99%.

Primitive-based contrast enhancement method

In this article we present a new solution for the problem of contrast enhancement. The method is based on the decomposition of the original image into its primitive components. In a first stage a multiresolution representation of the image is computed yielding transform coefficients proportional to the maximum of the local gradient at a specific scale and at a specific position in the image. As an image detail exists at several scales, the transform coefficients corresponding to the same image primitive are grouped. Contrast is enhanced by groupwise non-linear amplification of the transform coefficients. An inverse transform is then applied to the modified coefficients, resulting in a uniformly contrast-enhanced image without artifacts. Preliminary results demonstrate the effectiveness of this method. It could be used in a wide range of applications.