[Edaily Reporter Lee Hye-ra] #Dexter Studio has won an award at the longest-running international conference in the field of computer animation.

Dexter Studio announced on the 16th that a paper it presented on AI-based 3D motion search technology at “CASAXR 26,” held in Geneva, Switzerland, on the 15th (local time), won the “Best Applied Technology Award.”
CASAXR is an international forum that marks its 39th anniversary this year. It was founded in 1988 as the world’s first academic conference in the field of computer animation under the leadership of the Computer Graphics Society (CGS). In line with the rapid growth of extended reality (XR), the conference changed its name this year by adding “XR” to the existing “CASA,” transforming into a forum focused on future-oriented AI and immersive technologies.
Dexter presented on the topic of “Training-Free 3D Motion Retrieval from Video” (VMR). The key innovation is a technology that allows users to find desired 3D motions in a motion capture database by simply inputting a sample video, rather than searching via text combinations. The AI analyzes dynamic features to identify the most similar 3D motion data. This approach moves beyond the limitations of existing AI models, which require users to describe motions using subjective language, enabling intuitive exploration with just a single video containing the motion.
In performance evaluations using academic standard datasets, the technology achieved an accuracy of 99.66%, outperforming the text-based AI model (77.48%) developed by the Max Planck Institute for Intelligent Systems in Germany—a renowned institution that has established global standards in the field of virtual humans. According to the company, even when inputting videos from completely different visual domains—such as YouTube fan-cam footage, 2D animations, and AI-generated videos—the system excludes other elements and reads only the essence of the motion to produce precise search results with a similarity score of 0.899 or higher. It is reported that this enhances production efficiency by identifying similar 3D motion that can be immediately utilized in projects with only minor modifications. Another key feature is that it is a “training-free” AI model, meaning it can immediately apply newly updated open-source models without the need for separate system development, and upgrading to the latest model incurs no data retraining costs.
The paper for which Dexter received the award is scheduled to be published in the journal “CAVW” (Computer Animation and Virtual Worlds), published by Wiley, a global academic publisher with a 200-year history.
Lim Jae-ho, Deputy Director of Dexter’s R&D Institute, stated, “This is a technological achievement that simultaneously builds a modular pipeline capable of frequent updates using the latest open-source tools while securing unparalleled performance.” He added, “Leading-edge AI technology development not only enhances the efficiency of the production pipeline but also improves the utilization of Dexter’s high-quality database.”