Blog

  • Tiny-Face: Ultra-lightweight Face Detection for Mobile Devices

    Tiny-Face is an ultra-lightweight face detection model specifically designed to deliver fast and efficient performance on mobile and edge devices, where computational resources are limited. 🚀 Unlike many conventional face detection models, Tiny-Face is streamlined to use minimal memory and processing power while still achieving high precision in detecting faces.

  • RetinaFace: Single-stage Dense Face Localisation in the Wild

    Face detection in the wild presents unique challenges, but RetinaFace 🌐, a single-stage dense face localization model, tackles them effectively. My latest repository on GitHub explores this model, which enables high-precision facial detection and localization of key landmarks. This implementation integrates several backbone models (MobileNet and ResNet variants), allowing for flexibility between model size and accuracy. 🚀

  • Geometric Perspective of Vectors

    I have a strong background in mathematics, particularly in calculus from high school, and I can confidently perform various matrix multiplications and operations by hand. This has given me a solid understanding of vectors from both mathematical and computer science perspectives.

  • Real-Time Gaze Estimation Using Lightweight Models

    This project is focused on predicting the gaze direction using lightweight deep learning models. It combines classification and regression techniques to create a solution that is both efficient and accurate, making it ideal for real-time applications, especially on mobile devices.

  • Real-Time Head Pose Estimation with Powerful Backbones

    This project focuses on accurately estimating the head pose in real-time, and it’s packed with powerful features and improvements to ensure efficient and seamless performance. Whether it’s for AR/VR applications, attention tracking, or even driver monitoring systems, head pose estimation plays a crucial role in numerous industries.