Visual information retrieval techniques applied to veterinary radiology

This paper presents an assessment of the difficulties involved in implementing visual information retrieval (VIR) systems for veterinary images such as x-rays. Particularly, we examine cases where the x-ray exchange is of a casual nature, such as x-rays photographed and stored as JPEG images, so that they can be exchanged and shared more readily with other professionals in the field. While this practice is very convenient for the exchange of information between veterinary professionals, the resulting images contain imperfections such as glare, angular distortion, and background noise. These imperfections can become a stumbling block when applying VIR algorithms intended to analyze and process these images.