Daniel Zhang¹ and Jonathan Thamrun ², ¹ USA, ² California State Polytechnic University, USA

Injuries are common among both competitive and casual runners. Traditionally, patients seek medical experts like physiotherapists to treat their ailments. However, this can be expensive and time-consuming. To address this, we created an AI-driven mobile application that utilizes Natural Language Processing (NLP) and Computer Vision to diagnose users and recommend treatments based on user reported symptoms and uploaded images. This application was built using the Flutter framework and leverages Firebase for data storage. Additionally, it uses the OpenAI API to request OpenAI's gpt 4.1. In experimentation, we found that the model could detect injuries in knee X-Rays with about 64% accuracy and generate responses in around 82 seconds on average. These findings suggest that our method's accuracy is comparable to other methods, and that our response times are superior. Although our method isn't meant to replace medical experts, it can act as a swift, mobile first opinion for injured runners.

Machine Learning, Computer Vision, Nature Language Processing