[Haiqing Sub Forum] The latest achievements of computer vision research presented by two young scholars
 
time: 2017-05-11


Sub-forum of International Outstanding Young Scholar Forum has been held in B3-213 of School of Computer Science and Engineering of South China University of Technology on 11th May, 2017. Dr. Junle Wang from Ars Nova Systems and Dr. Jing Li from the University of Nantes in France shared their latest research results to teachers and students. The teachers and student representatives of the school attended the seminar.


Dr. Wang gave a seminar entitled From 2D to stereoscopic 3D visual attention-Psychophysical method and computational modeling. The seminar introduces how to use the algorithm to visualize the visual attention mechanism in the human visual system and extend it to 3D TV, VR and other 3D visual field. Dr. Wang revise the existing eye tracking technology to realize the three-dimensional image of the observer for eye movement data records and establish the first 3D eye movement database containing a variety of natural scenes. According to the eye movement data and images Depth information, a new 3D visual attention mechanism model is designed. The proposed model significantly increases the prediction rate in comparison of the existing methods. The model improves the coding of the 3D stereoscopic vision and the optimization of visual fatigue control.


The seminar presented by Dr. Li is Subjective Assessment methodology for multi-dimensional visual experience. It introduces the background of advanced multimedia technology such as 3D TV, VR and 4K TV and how to evaluate the multidimensional visual experience accurately. Based on the stereoscopic visual experience, Dr. Li modeled the behavior patterns of the viewers through the psychophysical experiments and proposed a new method of matching experiment. This method greatly improves the efficiency and accuracy of the subjective evaluation experiment of multidimensional visual experience. It also makes the interdisciplinary interdisciplinary subjective experiment cooperation and the network crowdsourcing experiment in the case of uncontrollable experiment environment. The method has been adopted by the International Telecommunication Union ITU and IEEE.P3333 as a standard for subjective evaluation of stereoscopic experiences.


After the meeting, teachers and students have discussion with the speakers. (Graphic / School of Computer Science and Engineering)