MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202621053119 A) filed by Deepali Suhas Jadhav; Maheshwari Virendra Ishvarshette; Payal Pandurang Tuptewar; Purva Balasaheb Patil; Ghansham Ashok Pawar; and Somanath Sidharam Birajdar on April 26, 2026, for A Mobilenet-Based Portable Ai System For Real-Time Bone Fracture Detection Using On Mobile Device Acquired Images Of X-Ray.
Inventors include Deepali Suhas Jadhav; Maheshwari Virendra Ishvarshette; Payal Pandurang Tuptewar; Purva Balasaheb Patil; Ghansham Ashok Pawar; and Somanath Sidharam Birajdar.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: Detection of fracture from radiograph images is crucial in medical practice and generally requires professional consultation, which can be unattainable in underprivileged areas or regions. In order to overcome such a challenge, we introduce in this letter a new solution for real time detection of fracture on mobile platforms running Android OS. Our method utilizes Convolutional Neural network (CNN) model, implemented with MobileNetV2 architecture, which is the optimal choice for solving the problem at hand both in terms of inference speed and accuracy of results. The model will learn to identify informative image characteristics, and thanks to the augmented x-rays used in training, it should be able to classify the images with fracture and without it accurately. After that, it will undergo optimization for TensorFlow Lite through quantization and compression in order to make the model more suitable for mobile inference. The model is encapsulated within an Android app, with capability of capturing photos using device camera or loading them from gallery, producing instant predictions with no Internet access required. Results demonstrate that the proposed solution is practical for CL applications, providing high accuracy and short inference time.
Disclaimer: Curated by HT Syndication.
