[Edaily Reporter Lee So-Hyun ] Team Naver demonstrated its expertise in vision AI by presenting 23 papers at the world’s leading academic conferences in the field of computer vision. A large number of research achievements related to spatial intelligence—including 3D spatial reconstruction, autonomous driving perception, world models, and image and video understanding—were accepted.

Naver (NAVER(035420)) announced on the 28th that 23 research papers from Team Naver’s technology organizations—including Naver Labs, Naver Labs Europe, Naver AI Lab, and Naver Cloud—have been accepted for publication at “European Conference on Computer Vision (ECCV) 2026.”

ECCV is recognized as one of the world’s most prestigious conferences in the field of computer vision, including video and image processing. Held every two years, this year’s conference will take place in Sweden from September 8 to 12. Team Naver plans to present the papers accepted for this conference.

At this conference, 23 Team Naver papers were accepted as regular papers across various vision technology fields, including 3D spatial perception and reconstruction, image and video learning models, and robot navigation simulation. In particular, the inclusion of numerous studies on 3D spatial reconstruction has been seen as a demonstration of the company’s competitive edge in spatial intelligence technology.

“BLASt3R,” the Successor to “DUSt3R,” Unveiled

Naver Labs Europe unveiled “BLASt3R,” the follow-up research to its 3D reconstruction technology “DUSt3R.” DUSt3R, a 3D reconstruction technology released by Naver Labs Europe in 2023, is an AI tool capable of reconstructing an entire space in 3D using just one or two photos.

The newly unveiled BLASt3R incorporates Simultaneous Localization and Mapping (SLAM) technology, which enables a robot to create a map of its surroundings in real time while navigating. This allows for 3D reconstruction even in on-device environments and improves both speed and accuracy compared to previous models.

Naver Labs is also presenting research on indoor 3D space generation. This research involves technology that reconstructs a room’s structure and objects using a single 360-degree panoramic image.

Autonomous driving perception technology was also included. Naver Labs unveiled a technology that utilizes a Vision-Language Model (VLM) to recognize even objects not previously learned during autonomous driving and reflect them in real time on a bird’s-eye view (BEV) map.

Four Studies, Including ‘Seoul World Model,’ Selected for Spotlight

The “Seoul World Model” research, conducted by Naver AI Lab in collaboration with KAIST, was also accepted as a formal paper. The Seoul World Model is a technology that uses spatial data held by Naver—such as panoramic images of the entire city—to generate long-range videos depicting the actual appearance of Seoul.

This research has been recognized for its potential applications in various fields, including map-based virtual exploration, digital twins, and autonomous driving and robotics simulations.

Naver Cloud is unveiling research achievements that enhance AI’s ability to understand images and videos. Studies accepted include research aimed at accurately recognizing the movement of objects within videos, as well as research that adds text-based inference to spatial images to make it easier to perceive depth.

Among the accepted papers, a total of four—including research on 360-degree spatial reconstruction and the Seoul World Model—were selected for the “ECCV Spotlight.” Spotlight status is awarded to approximately 1–2% of all submissions.

Team Naver has continued to achieve double-digit paper acceptance rates every year at the three major conferences in the field of computer vision—including ECCV, the Computer Vision and Pattern Recognition Conference (CVPR), and the International Conference on Computer Vision (ICCV).

A Naver spokesperson stated, “We plan to continue our research and development of next-generation AI and spatial intelligence technologies and integrate them into various services to provide users with tangible technological experiences.”