Karma AI is an assistive multimodal AI system developed to provide real-time contextual assistance to visually impaired individuals. It utilizes state-of-the-art technologies, integrating voice commands, advanced image analysis, and natural language processing powered by Google Cloud's Vertex AI and OpenAI, to deliver accessible audio feedback about the user's surroundings.
- Voice Interaction: Users communicate naturally via speech, initiating tasks and receiving clear audio responses.
- Real-time Image Recognition: Leverages Google Cloud Vision API and ML Kit for rapid object detection and scene understanding.
- Emergency Fall Detection: Built-in accelerometer-based detection system identifies falls, automatically triggering emergency calls and messages if necessary.
- AI Scene Description: Captures and interprets images in real-time using Google's Gemini API, providing succinct verbal descriptions tailored for navigation and safety.
- ChatGPT Assistant: Integrated GPT-4o-mini model offers concise, helpful guidance through conversational interactions, specifically optimized for visually impaired users.
- Google Cloud Vertex AI, Gemini Pro Vision API
- OpenAI GPT-4o-mini
- Android (CameraX, TTS, Speech Recognition)
- Kotlin, ML Kit, Firebase
🏆 Social Impact Award – 2025 Applied AI Challenge
Hosted by the Student Innovation Center at Iowa State University
Awarded to Karma AI for effectively addressing accessibility challenges using advanced multimodal AI technologies.
- Rahman Abdul Rafi
- Ankit Jyothish
- Kunal Suresh
- Arun Kumar Iyyappan
- Mohammed Musthafa Rafi