Medisur (Oct 2022)
Radiographic Measurements and Diagnosis of Hip Dysplasia in Infants Using Computer Vision and a Rule-Based System
Abstract
Background: the use of digital radiographs for the diagnosis of developmental dysplasia of the hip allows, in addition to early diagnosis and greater work efficiency, more precise measurements, patient monitoring and surgical planning.Objective: to describe a tool capable of detecting structures and points of interest in a semi-automatic way to carry out the necessary measurements and calculations with a view to diagnosing hip dysplasia in infants.Methods: study of technological innovation, where the Viola-Jones artificial vision algorithm was used for the detection of structures, as well as a rule-based system with a view to a diagnosis suggestion. The proposed tool (Software for radiographic measurements for diagnosing developmental dysplasia of the hip in infants) is in the testing and exploitation phase at the Paquito González Cueto University Pediatric Hospital, in Cienfuegos. To validate the results, radiographic studies of 12 cases were chosen, to which measurements were applied using the traditional method and then using the software.Results: a system was obtained with a view to determining structures in hip radiography images, which allow points and lines to be obtained to calculate dysplasia indicators. The success rate in detecting the structures was 100%.Conclusion: there was a high coincidence between the measures calculated by the algorithm and those calculated manually. The correspondence between the diagnosis predicted by the system and that issued by specialist doctors was also high.
