This is an online seminar. Registration is required.
【Robotics for Infrastructure Management Team】
【Date】2021/May/21(Fri) JST1420pm-1520pm
Title: Machine learning for Multimedia Communications and Computer Vision
Abstract: Multimedia (e.g., videos and images) is one of the most important content types on the Internet nowadays. However, there are still many challenges in both transmission and understanding of multimedia content. In this seminar, I will present our recent work on machine learning-based methods for addressing various important problems in multimedia communication and computer vision. Regarding multimedia communication, we developed machine learning-based high-performance Quality of Experience (QoE) models for not only traditional videos but also Virtual Reality (VR) videos. Related to computer vision, we conducted for the first time an evaluation of human action recognition algorithms considering the influence of video transmission and adversarial perturbations. In addition, we are currently developing a deep learning-based cloud detection algorithm for ground-based whole sky images to facilitate satellite laser communications.
