Over the past five years, Naver has successfully published 151 papers in the world’s top three computer vision conferences, including CVPR, ECCV, and ICCV. Recently, Naver announced that 14 research papers from its team, which includes Naver Labs, Naver Labs Europe, and Naver Cloud, have been accepted to the 2025 CVPR. The CVPR conference, co-hosted by IEEE and CVF since 1983, is recognized as a leading platform in vision AI and a highly influential one in computer science.
The Naver team plans to showcase follow-up research on ‘DUSt3R,’ a 3D reconstruction AI tool introduced by Naver Labs Europe that received significant attention from global tech firms. This includes new AI models like ‘MUSt3R’ for reconstructing more accurate 3D spaces from multiple images and ‘Pow3R’ for enhanced inference capabilities by integrating various camera and scene information.
Additionally, several papers on spatial intelligence, including a novel object pose estimation method by Naver Labs, were accepted. The rise of multimodal AI technologies has also brought attention to related image learning model research, further cementing Naver’s presence in the field.
Naver Cloud proposed an efficient learning methodology using masking techniques to address training instability in image learning models. Meanwhile, Naver Labs Europe demonstrated the ability to accurately differentiate new objects in images using vision-and-language models without additional training.
In robotics, Naver showcased an end-to-end self-driving system for rapidly navigating robots and proposed methods to enhance creativity in image generation AI models.
Overall, Team Naver has published over 450 research papers in top AI conferences in various fields, including natural language processing and machine learning, with their work being cited over 47,000 times. A Naver representative emphasized continued investment in R&D to secure advanced technologies, aiming to integrate their key innovations into Naver’s main services for a distinctive user experience under their “On-Service AI” strategy.