🌊 Online machine learning in Python

Real-time GCC-NMF Blind Speech Separation and Enhancement

The World's First Realtime onDevice AI Project to maintian Modestey and Religious values. Everyday AI companion designed for Muslim community to empower them maintain modesty, decency in the digital realm by blurring and covering Non-Mahram and Haram objects (NSFW), directly on screen on anywhere(offline,online,browser,videoplayer) in anytime.

An open source, cloud-native streaming data integration and analytics product optimized for agile digital businesses

(e2e) foss trading for non-tinas

Project for real-time anomaly detection using Kafka and python

An intuitive GUI for simplifying attendance tracking and management with Face recognition

Revisiting K-Net for Real-Time Panoptic Segmentation. Code release for our IV 2023 paper.

This repository houses a full-stack application which is used to create a real time stock price dashboard in Python, with the use of the streamlit and yfinance libraries.

Toolbar application that leverages OpenCV's powerful computer vision capabilities and MediaPipe's pose detection to create an intelligent posture monitoring system. Running discreetly in your system tray, it provides real-time feedback and gentle reminders when your posture needs attention.

This project aims to improve the accuracy of blood pressure measurements using video-based heart rate monitoring and a calculation algorithm. The system utilizes facial tracking to detect and analyze heart rate data, which is then used to estimate blood pressure.

Microservices Framework for Stopover

Real-time driver drowsiness detection system using computer vision, OpenCV, and MediaPipe face mesh. Monitors eye movements and calculates Eye Aspect Ratio (EAR) to detect fatigue, providing immediate visual alerts for enhanced driving safety.

This is a project which demonstrates creation of a data pipeline by scraping data using twitter API and creating a data delivery stream using Kinesis Firehose for ingesting data to Amazon S3.

An OpenCV-based AR pass-through headset with overlay for interesting new ways to see the world

A real-time voice-to-text and text-to-speech AI pipeline using Whisper, an LLM, and Edge-TTS with tunable parameters for low-latency audio processing and response generation.

A web application for real-time pose estimation and exercise form correction using OpenCV, MediaPipe, and Flask. Supports various gym exercises like squats, push-ups, bench press, and more.

Real-time object detection (YOLOv8), tracking (SORT), and text recognition (EasyOCR) pipeline deployed on AWS using producer-consumer architecture with Kinesis Video Streams, EC2, SQS, Lambda, S3, Textract and DynamoDB.

A custom-trained YOLOv11-based license plate detection with PaddleOCR for accurate ANPR.

The project focuses on extracting and structuring key details from resumes, such as names, contact information, education, and work experience, into a user-friendly interface for efficient review and management.