Until now, it has required great effort for filmmakers to insert animated characters in feature film scenes: In the so-called motion capture process, real actors wear skintight suits with markers. The suits reflect infrared light, which is emitted and picked up by special cameras. Subsequently, the movements of the actors are transferred onto animated characters via software.
In the future, this process could be facilitated by a method developed by Chenglei Wu from the Max Planck Institute for Informatics in Saarbrücken. The PhD student is working on a new approach to transform scenes of images or videos into detailed and flexible three-dimensional models without great effort. "This is even possible with videos that are taken in uncontrollable situations, like varying lighting conditions in natural surroundings", explains Wu. However, the technique he uses to provide images and videos with spatial dimensions is relatively easy: only one or two cameras are needed. They film the movements of the actors. An algorithm converts them and transfers them into three-dimensional models of the characters before they are put into the movie. Moreover, this technique is able to treat the webcam video data in such a way that an extremely precise and highly flexible face model is created. Nevertheless, the video does not need to be available in its best quality. The approach is not only interesting for filmmakers and photographers, but also for physicians: The technology offers them a better way to analyze movement sequences or healing processes. Moreover, material testers could use the method to examine deformable surfaces.
"I am delighted about the award. It is a great honor for Chenglei Wu", says Christian Theobalt from the Max Planck Institute for Informatics. "It shows that the Graduate School of Computer Science in Saarbrücken is an institution for excellent students from all over the world to work on their doctoral degree".
Chenglei Wu did his master's degree in computer science at the renowned Tsinghua University in Beijing. He started working on 3D models back in China. Within the context of his PhD in Saarbrücken at Saarland University, he delved further into the topic. As a doctoral candidate, Wu is a member of the Graduate School of Computer Science in Saarbrücken, which promotes its PhD students with a structured program and scholarships. Moreover, the Intel Visual Computing Institute of Saarland University supported the researcher in the context of his dissertation. Wu will finish his PhD in the coming months. After that, he will continue to work on his research at the Eidgenössische Technische Hochschule (ETH) in Zürich.
The "Chinese Government Award for Outstanding Students Abroad" is tendered by the Chinese government every year. It goes to 500 excellent Chinese doctoral candidates who are working abroad on their academic careers in different disciplines. Besides Wu, another computer science PhD student in Munich was honored with the award.
The award ceremony took place on the 24th of May at the Chinese embassy in Berlin.
Background information about computer science research at Saarland University in Saarbrücken
The Department of Computer Science represents the center of the computer science research in Saarbrücken. Seven other world-renowned research institutes are close by the department: The Max Planck Institutes for Informatics and for Software Systems, the German Research Center for Artificial Intelligence (DFKI), the Center for Bioinformatics, the Intel Visual Computing Institute, the Center for IT Security, Privacy and Accountability (CISPA) and the Cluster of Excellence "Multimodal Computing and Interaction".
Videos that show how characters are reconstructed for movie scenes or portrait photos are available here:
Questions can be addressed to:
Prof. Dr. Christian Theobalt
Graphics, Vision and Video
Max Planck Institute for Informatics
Phone: +49 681 / 9325 4028
Competence Center Computer Science Saarland
Phone: +49 681 /302 70741
AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert! system.