The correct interpretation of visual information is an important factor in many applied settings. The analysis of visual search behaviour of human observers within these situations can give a unique insight into the perceptual and cognitive processes being used. Often the skills that observers develop are complex and only arise from many years of exposure to a particular task and consequently the nature of these skills is not easily investigated. A greater characterisation of the skills required for expert performance within a domain is essential if the number of errors that are made is to be minimised. This presentation compares the findings from two different applied settings that rely heavily on visual inspection, namely; medical imaging and airport baggage inspection.

For a number of years we have been investigating how expert radiologists interpret the information contained in radiographic (X-ray) images, which are two dimensional representations of three dimensional anatomy. Consequently the pathologies that can be visualised on a radiograph are often obscured by overlying structures, making their correct detection and interpretation difficult. Also within medical screening environments (such as breast cancer screening) the radiologist must examine many normal images in order to detect the very few abnormal variants that may indicate the presence of a malignant process.

The X-ray screening of passenger baggage is carried out by trained security personnel. Each item of luggage is individually examined to detect a variety of restricted items that may compromise the security of the aircraft. These items include gas canisters, knives, guns and explosive devices. As in a medical screening environment, the vast majority of images that are examined will contain no restricted items, but it is vital the few bags that do contain these target items are correctly identified.

The results of a number of studies that we have carried out examining both medical image and baggage X-ray interpretation are presented. The use of eye movement recording equipment throughout these studies has enabled different aspects of the skills involved to be identified. Also tachistoscopic image presentation has been used to examine the relative importance of focal and peripheral visual processing. Data from the studies allow a comparison between the two areas across a number of different visual search parameters. These include fixation and saccade characteristics, general search patterns, time to locate targets, and image coverage. The relationship between image characteristics and visual search guidance is investigated.

The results of the studies show that there are similar patterns of visual search behaviour between the two domains, but there are also very important differences. The nature of these differences and the implications for procedures and training within the two domains are examined. This research demonstrates that monitoring visual search within two different applied environments can give unique and valuable insights into the visual inspection process. The studies also demonstrate that care must be taken when attempting to extrapolate the results between different applied settings even when they appear to be very similar.