Block-Deep: A Hybrid Secure Data Storage and Diagnosis Model for Bone Fracture Identification of Athlete From X-Ray and MRI Images

Abstract

A human adult’s skeleton is made up of 206 bones that perform a variety of vital biological tasks like safeguarding the internal organs and preserving vital nutrients. Bone fractures in adult’s particularly in athletes can lead to minor sports injuries or even potentially fatal ones for which a sports injury specialist may be needed for treatment. The traditional methods of bone fracture detection are costly and time consuming and may leads to in-accurate results due to fault in the diagnostic machines. On the other hand, machine learning methods have found widespread use in bone fracture detection using X-ray or MRI images. However, all such methods haven’t any proper security mechanism to athlete data that leads to theft, fraud and misuse of sensitive information. Therefore, with the correct fracture detection processes there also exists a strong security measure to stop unauthorized people from accessing or changing the data. In this research a hybrid model based on blockchain and deep learning has been proposed to diagnose the bone fractures as well as protection of the confidential data of the athletes. The key steps of the proposed work are data collection, blockchain based data security and transformation, feature extraction through Capsule Network, final classification using Visual Transformer based transfer learning. From the experimental evaluation and comparison with the state-of-the-art methods, it has been observed that the performance of the proposed work is excellent in terms of accuracy, with the value of 95.01%, 94.04 and 96.25% on different dataset respectively.

Keywords

• Bone fracture
• blockchain
• capsule network
• MRI
• machine learning
• X-ray