Bone Fracture (BF) is common and could lead to malunion and arthritis with attendant morbidity (Jones et al., 2020), mortality and around 1.71 billion people Musculoskeletal Condition (MC) worldwide (Cieza et al., 2020). Besides, older adults aged ≥65 has with MC have a greater chance of getting a Musculoskeletal Fracture (MA). Proximal femur fractures, including osteoporotic femoral neck fractures, are also very common to this group, and the rate is expected to be doubled in the next 30 years. Additionally, other fractures like oblique, compound, comminuted, spiral, greenstick and transverse are also seen in the BF and MF domain.
Therefore, diagnosing and treating a BF patient is essential. On the contrary, missed fractures are a common diagnosis error in the Emergency Departments (EDs), and ~1% of the total fractures are missed by the clinicians. This leads to treatment delays and complications in healthcare settings. Additionally, many patients' perception of fracture information and clinician's overworked schedule may also add extra complications to therapeutic decision-making and further delay surgical repair. Although, Magnetic Resonance Imaging (MRI), computed tomography (CT), Nuclear Medicine Bone (NMB) scan, Ultrasound and X-ray are used to identify the fracture, except Ultrasound and X-ray, other forms of technology are expensive and are not easily accessible. Furthermore, there is a lack of automated fracture detection tools and on the contrary, CT and X-ray scan image-based manual orbital fracture detection leads to various challenges.
Our priority for this research is to understand the ML settings and find out the research gap for the researcher to understand the best techniques to recognise Bone fracture. We have conducted a systematic literature review by finding each of the paper's Categories, motivations, the problem they have addressed, solutions they have proposed, novelty, limitations, image quantity, the models and finally, the accuracy of each of them. Furthermore, we have analysed and compared them by their accuracy based on the image quantity, percentage of models used, practical implementation, limitations, research progress, and accuracy based on image types.
Hossain, E. (2021, November 29). Eliashossain001/bone-fracture-detection: Detecting fractures on top of X-ray imaging modalities using various state-of-the-art techniques of Deep Learning. Retrieved [Put your current date], from https://github.com/eliashossain001/Bone-Fracture-Detection