Speaker/Presentation

Time in CEST

Prof. Dr. Yun Zhang

Chair Introduction

9:00 - 9:05 am

Prof. Dr. Hanli Wang

Visual Translation: From Image and Video to Language

Abstract:
Translating an image or a video automatically into natural language is an interesting, promising, but challenging process. The task is to summarize the visual content of the image or video and to re-express it with the correct words and suitable grammar, sentence patterns, and human habits. Nowadays, the encoding–decoding pipeline is the most commonly used framework that is implemented to achieve this goal. In particular, convolutional neural networks are used as the encoder to extract the semantics of images or videos, while recurrent neural networks are employed as the decoder to generate word sequenced. In this talk, the literature on image and video description is first reviewed, and preliminary research advances, including visual captioning, visual storytelling, visual dense captioning, visual sentiment captioning, and more complex visual paragraph description, are introduced.

9:05 - 9:50 am

Q&A

9:50 - 10:00 am

Dr. Junhui Hou

Deep Learning-based 3D Point Cloud Reconstruction

Abstract:
Three-dimensional point clouds are widely used in immersive telepresence, cultural heritage reconstruction, geophysical information systems, autonomous driving, and virtual/augmented reality. Despite the rapid developments in 3D sensing technology, it is still time consuming, difficult, and costly to acquire 3D point cloud data with a high spatial and temporal resolution and complex geometry and topology. In this talk, I will present recent studies on computational methods (i.e., deep learning)-based 3D point cloud reconstruction, including sparse 3D point cloud upsampling, the temporal interpolation of dynamic 3D point cloud sequences, and adversarial 3D point cloud generation.

10:00 - 10:45 am

Q&A

10:45 - 10:55 am

Closing of Webinar
Prof. Dr. Yun Zhang

10:55 - 11:00 am