| Autor: |
Narayanan, Ramkumar, Rangan, Venkat, Gopalakrishnan, Uma, Hariharan, Balaji |
| Předmět: |
|
| Zdroj: |
Proceedings of the International Conference on e-Learning (ICEL); 2016, p190-197, 8p |
| Abstrakt: |
Current e-Learning technologies enable a multitude of distant students to simultaneously attend a live teaching session conducted by a remote instructor. The instructor views the video streams of students from multiple locations on displays surrounding him. However, when there are an increased number of distant students in each location, the instructor finds it difficult to view and interact with each of them. Overtime, he feels a lack of connectedness with the poorly visible distant students. In this paper, we have implemented a system that provides the instructor with an immersive, zoomed in view of the student that the instructor is currently interacting with. In addition to that, the student details are overlaid on the video. The system involves real-time gesture recognition for identification of interacting students using Microsoft Kinect and a mapping of the coordinates of the interacting students from Kinect to the pan-tilt-zoom camera's frame of reference. We employ techniques such as the Viola Jones algorithm for face detection and the Optical-Flow algorithm for tracking movements during pan, tilt and zoom. In addition, a face recognition system with an overlay of student details beside the students face enables the teacher to tailor the interaction according to the knowledge background of the student. Subjective analysis indicates that the instructor can interact with the remote students more effectively and with ease. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
