Junior Seminar

Mercoledì 6 marzo alle ore 16, Muhammad Usman (dottorando, Università degli Studi Roma Tre) terrà il seminario dal titolo "3D Human Pose Estimation: Real-time Performance and Applications".

Abstract:
Human pose estimation is a task that involves identifying the location of specific points in an image, usually referred to as keypoints. The keypoints can represent various parts of the object such as joints, landmarks, or other distinctive features. YOLO (You Only Look Once), a current state-of-the-art deep learning object detection model, simultaneously identifies and predicts the keypoints that define the human body's pose. In our talk, we will present a real-time 3D human pose estimation system by employing two cameras and triangulation methods to calculate the 3D pose from the 2D pose predicted by YOLO. Our results showcase the potential of the proposed method as a robust tool for real-time 3D human pose estimation. Beyond accuracy and efficiency, this approach opens up numerous possibilities across various domains, including human-computer interaction, sports analytics, and surveillance.

Il seminario si svolgerà in presenza presso il Dipartimento di Matematica e Fisica, via della Vasca Navale, 84- aula 108.