Deep Learning Analysis of Bone Fracture Using Images Processing Techniques

Abstract

Fracture can be defined as a condition of breakage or lack of bone continuity. Computer-based techniques are increasingly being used to identify faults and cracks. An effective system can be identified by two important characteristics: it must be able to detect quickly and accurately, while utilizing modern techniques and utilizing resources efficiently. Bone fractures are caused by an excess external force that exceeds the bone's threshold. A Canny Edge detection method uses automated fracture detection to detect bone fractures, which overcomes the noise issue. Several edge detection techniques are available., including Canny, Log, Prewitt, and Robert. Because these techniques cannot perform multiresolution analysis, they are not useful for detecting minor details during analysis. Additionally, when dealing with noisy images, these techniques cannot work as well as they do with high-quality, high-resolution images because they are not able to differentiate edges from noise. To overcome this this problematic condition, we proposed deep learning based Convolution Neural Network (CNN). We discovered that the suggested strategy offered superior edge identification results at aggregate levels when taking our simulations into account. The suggested approach has been shown to be more resilient than the present edge detectors in terms of extracting the required data, performing the required processing, and handling noise.