Detection of tibial fractures in cats and dogs with deep learning

The aim of this study is to classify tibia (fracture/no fracture) on whole/partial body digital images cats and dogs, localize the fracture by using deep learning methods. This provides diagnose more accurately, quickly safe for clinicians. In study, a total 1488 dog cat that were obtained from universities institutions used. Three different studies implemented detect tibia. first phase was classified automatically as or no with Mask R-CNN. second phase, location in image localized directly third SSD. accuracy F1 score values 74% 85%, respectively value 84.5%. 52.1% 68.5%, respectively. 46.2%. results research showed promising detection fractures dissemination diagnosis help such smart systems would also be beneficial animal welfare.