Artificial Intelligence Based Realtime Face Tracking Gimbal for Live Video Session

Abstract: The Real-Time Face Tracking Gimbal for Live Video Session is a dynamic system that enhances video streaming by automatically tracking and cantering a subject’s face in the camera frame. Utilizing advanced algorithms for face detection and tracking, this system ensures smooth camera movements across pan, tilt, and roll axes. It leverages a combination of image processing and embedded systems to achieve real-time performance. The gimbal’s camera captures video, processed by a microcontroller integrated with a pre-trained model, which detects the subject's face and calculates its position within the frame. Any deviation prompts the microcontroller to adjust the gimbal's motors, ensuring the subject remains cantered. This closedloop feedback system makes it ideal for live streaming, vlogging, and professional videography. The system’s design focuses on portability, scalability, and ease of use, incorporating hardware such as servo motors, motor drivers, and a USB (Universal Serial Bus) camera. It also supports software tools for live streaming and algorithm customization. Applications span entertainment, education, and virtual meetings. By automating face tracking, the gimbal reduces manual intervention, improves video quality, and ensures a seamless viewing experience for audiences. This design reduces the need for manual camera handling, offering a more professional and user-friendly experience. Its compact, portable structure makes it suitable for live streaming, vlogging, video conferencing, and surveillance applications. Furthermore, the system is highly scalable and costeffective, allowing for integration into various industries requiring real-time tracking