The problem OpenMed solves

Access to Medical Diagnoses: In many parts of the world, including India, there is a shortage of healthcare professionals, especially in rural areas. This platform provides easy and unlimited access to medical diagnoses, bridging the gap between people and healthcare services.

Personalized Healthcare Advice: The platform behaves like a family doctor, offering personalized healthcare advice based on individual medical histories and symptoms. This helps users make informed decisions about their health.

User Anonymity: Maintaining user anonymity is a significant concern in healthcare, as people may hesitate to seek medical advice if they fear their privacy will be compromised. The platform addresses this issue by providing a confidential and anonymous environment for users.

Language Barriers: By supporting local Indian languages, the platform makes healthcare information more accessible to a wider population, including those who may not be proficient in English.

Health Data for Public Health: The platform contributes to public health monitoring by providing relevant data to government and concerned authorities, helping them identify and respond to possible outbreaks or health trends in specific regions.

Blockchain for Identity: The use of blockchain technology for user identification and anonymity adds an extra layer of security and privacy to the platform, making it more trustworthy for users.

Challenges we ran into

One of the primary challenges our team encountered while developing the healthcare platform was ensuring robust Speech-to-Text (STT) and Text-to-Speech (TTS) support for local Indian languages. This challenge is significant because it directly impacts the platform's accessibility and usability, particularly for users from diverse linguistic backgrounds across different regions of India.

To address this challenge, we explored various solutions, both locally hosted and cloud-based, in an effort to find the most suitable STT and TTS models. This process involved significant research, experimentation, and testing to identify the best-performing technology. After thorough evaluation, we settled on using Whisper, an open-source vec2word model.

Whisper, with its versatile capabilities and support for local languages, emerged as a robust choice for handling the linguistic diversity of India. This decision not only enhances the accessibility of our platform but also ensures that users can comfortably converse with the AI in their preferred regional language.

Another noteworthy challenge was integrating TTS audio with an Augmented Reality/Virtual Reality (AR-VR) model. This integration posed unique technical difficulties, especially when considering the complexities of synchronizing audio output with AR-VR experiences.

Additionally, we encountered challenges related to Azure access subscription expiration, which impacted the development process. Azure services played a crucial role in our platform's functionality, and this interruption in access necessitated adjustments in our development workflow.