Hip dysplasia (HD) in dogs is a hereditary disease, with a polygenic, additive nature and with high prevalence in specific breeds. Without ideal treatment, the diagnosis of this disease aims to improve the life quality of dysplastic animals and support the reproductive selection process.
Due to the large number of genes involved and the complexity of genetic transmission, radiographic diagnosis is the only way to detect whether or not dogs are affected by this disease. The diagnosis of HD is recommended in healthy (screening) and sick animals. Radiographic examination can be performed on younger animals using a stress radiograph in which hip laxity, the major risk factor for the development of hip dysplasia, is evaluated by calculating the Distraction Index (ID), or in more adult animals, by means of a radiograph in which it is possible to determinate the Norberg angle, which can be translated into a scale of different degrees of HD. The animal is positioned on the radiographic examination table by one or more veterinarians to obtain the appropriate technical quality. Currently, the reading of the radiography, and consequent diagnosis of hip dysplasia is a manual, subjective and very demanding process in terms of human resources, time and costs.
To make the process faster, more objective and accurate, the main objective of this project is to develop a technological platform to support the diagnosis of hip dysplasia in dogs. Therefore, based on computer vision techniques and machine learning models (Machine and Deep Learning), it will be possible to streamline and automate the process of radiographic analysis and to create a software capable of supporting the clinical decision in terms of radiographic reading, HD classification and certification. The aim is to create a CAD system capable of assessing the technical quality of the images (regardless of the type of equipment used) in terms of positioning and to classify the radiographs for both young and adult dogs.
NEADVANCE – MACHINE VISION, S.A.
HOSPITAL VETERINÁRIO MONTENEGRO