This project integrates deep learning, machine learning, and natural language processing to provide a comprehensive AI driven Vision Care system. The CNN based analysis enables accurate detection of retinal abnormalities through image processing, while the RFC prediction model offers data driven insights based on numerical parameters. Additionally, the LLM chatbot enhances user engagement by providing interactive consultations and personalized recommendations. By combining these three approaches, this tool aims to assist individuals in monitoring their eye health efficiently, promoting early detection, and encouraging preventive care.
This project detects driver drowsiness using EMG (Electromyography) signals captured through the BioAmp EXG Pill sensor connected to an Arduino. Real-time data is visualized using Python, and alerts are generated (beep sounds) when muscle activity drops below a defined threshold combined with eye and face data from Raspberry camera module — indicating possible fatigue or drowsiness, this is done using Open CV. Everything is managed via an app that also predicts real-time location where the driver should rest to prevent fatigue.
Tarang AI is proudly rooted in India and is backed by real data of 1700+ places. This project predicts water safety based on environmental parameters using a Random Forest Classifier model. It preprocesses water quality data, trains a Random Forest Classifier model, calibrates it, and provides predictions i.e. probability of the water being safe or probability or it being unsafe. It uses advanced AI/ML algorithm to predict water safety across based on real-time inputs from the user. It also offers other features like water safety tips, flagging personal concerns etc.
Yuri Cloud uses machine learning techniques like isolation forest to flag potential threats in your cloud infrastructure. This project was designed to enhance cloud security through real-time monitoring and AI-driven anomaly detection, it helps in alerting the user of potential threats and performance issues. YuriCloud features user-friendly interface with a dashboard that displays basic information about the cloud server, an explore page that provides cybersecurity news and other related articles to keep the users informed and alert. The main feature it offers is to perform real-time check and get a full graphical report on potential anomalies.
This project implements a real-time voice-based AI assistant using LiveKit's Agent framework. It integrates automatic speech recognition (ASR), natural language processing (LLM), speech-to-text (STT) and text-to-speech (TTS) using APIs from Deepgram, GROQ, and Cartesia, along with multilingual voice activity detection and noise cancellation. The agent works locally on your console and delivers impressively low latency across all stages of interaction in real time, from detecting when the user finishes speaking to delivering a fully spoken response. During each call session, detailed performance metrics such as End of Utterance Delay (EOU), Time to First Token (TTFT), Time to First Byte (TTFB), total latency, etc. are stored automatically in an excel file.
