Bone fracture dataset
Description
This dataset is prepared for bone fracture classification for tibia and fibula bones. The data format is PNG, which actually proved by medical experts from the University of Gondar, Ethiopia, and some was taken from an online repository. This dataset is intended for classifying bone fractures in the tibia and fibula bones. This dataset is prepared for bone fracture classification for tibia and fibula bones. The data format is PNG, which actually proved by medical experts from the University of Gondar, Ethiopia, and some was taken from an online repository. The goal is to develop a machine learning model that can accurately classify and localize bone fractures from the provided images. This dataset will serve as a valuable resource for training and testing such models, ultimately contributing to the improvement of fracture diagnosis and treatment
Files
Steps to reproduce
The study was conducted following the design science approach. The data were gathered from the University of Gondar referral hospital and the MURA (musculoskeletal radiographs) repository. After the data, the image quality improved using median, gaussian, mean, and max filtering and was checked with the help of image quality metrics PSNR, MSE, SSIM, and cosine similarity. After enhancement, image augmentation and segmentation techniques are applied. Hereafter, features are extracted by CNN. We employed different classifiers (SVM, RF, and AdaBoost) with a training and testing dataset split ratio of 80/20.