開催日時：2024年4月5日(金) 10:30 ~　90分

場所：講堂多目的室１　および　Web併用（※対象は特に限定しない）

講演者：Hongjin Ren先生（Nanjing University of Science and Technology）

題目： Exploring Feature Fusion Methods for Enhancing High-Dimensional Images and Videos

概要：
With the rapid development of digital technology and the explosive growth of big data, there is a growing demand for higher-quality videos/images as carriers of information transmission. However, many scenes or early records cannot meet the modern high standards of information carrying due to limitations of shooting equipment. Therefore, researchers widely employ super-resolution technology to magnify and enhance the details of these images and videos to a certain degree. In the context of deep learning-based super-resolution tasks, the fusion and interaction of high-dimensional image and video features have received significant attention. This issue involves effectively integrating features from different dimensions, which has a crucial impact on the performance of models. This presentation aims to use stereoscopic videos and hyperspectral images as examples to introduce the current mainstream methods and latest advancements in super-resolution. Additionally, we will explore the potential of integrating and cross-applying these methods. Furthermore, we will combine our research cases to conduct specific analyses, aiming to achieve sufficient fusion and effective interaction of features from different dimensions in videos/high-dimensional images at minimal cost, thereby improving model performance and efficiency.

お問い合わせは以下のアドレスにメールでご連絡ください。
dser-center@biwako.shiga-u.ac.jp