• home
  • News Hub
  • Hanwha Vision Unveils Breakthrough Deblurring Technology at ECCV

2024.10.02

Hanwha Vision Unveils Breakthrough Deblurring Technology at ECCV

□ “Deblurring 3D Gaussian Splatting” to revolutionize video surveillance with faster and sharper image restoration

□ Hanwha Vision’s support for talent drives technological advancements

A scene reconstructed using 3D Gaussian splitting (left) and restored using deblurring techniques (right).

Hanwha Vision, a global vision solution provider, recently presented its deep-learning-based deblurring research at the European Conference on Computer Vision (ECCV) in Milan, Italy, starting September 29th.

Held every two years, ECCV is recognized as one of the top-tier computer vision conferences globally, ranking alongside CVPR (Computer Vision and Pattern Recognition) and ICCV (International Conference on Computer Vision).

Outperforming Existing Deblurring Methods

During ECCV, Hanwha Vision showcased its latest research, demonstrating significant improvements in the clarity and detail of blurred images.

Hanwha Vision has introduced the first deblurring technique applied to 3D Gaussian Splatting, a technology widely used in virtual and augmented reality. By integrating multi-layer perceptrons (MLP) along with other advanced components, Hanwha Vision has overcome the limitations of traditional 3D Gaussian Splatting in handling blurred images, resulting in a substantial boost in image sharpness and processing speed.

This innovative model surpasses other existing state-of-the-art techniques, including those based on NeRF (Neural Radiance Fields), a core technology for the metaverse. It delivers superior visual quality while further accelerating rendering speeds.

“This research marks a major milestone by addressing the limitations of 3D Gaussian Splatting while enhancing its strengths through deblurring technology,” said Jeongeun Lim, Head of Hanwha Vision AI Lab. “It opens new possibilities in image restoration. Hanwha Vision will keep pushing the boundaries in video surveillance and computer vision through proactive research and development.”

Nurturing Talent and Fostering Innovation

Led by Howoong Lee, a senior research engineer of Hanwha Vision AI Lab, the study also highlights Hanwha Vision’s commitment to nurturing top AI talent and supporting employee development.

A graduate of Sungkyunkwan University, Lee benefited from the company’s Science and Engineering Scholarship Program, designed to cultivate experts in the rapidly expanding AI sector. This program provides exceptional employees with unique opportunities to become AI specialists, fostering positive experiences, loyalty, and commitment within the organization.

Lee shared, “Thanks to the company’s support, I was able to explore cutting-edge technology and achieve innovative results. With Hanwha Vision behind me, I look forward to continuing my research and contributing to further technological advancements.”

Hanwha Vision‘s commitment to employee development extends beyond scholarships. The company’s innovative ‘Community of Practice (CoP)’ initiative fosters a collaborative environment where employees of all levels can share knowledge and learn from each other. This culture of collaboration and continuous learning encourages employees to explore new ideas and technologies, making Hanwha Vision a truly inspiring place to work.

Furthermore, since establishing its AI Lab in 2019, Hanwha Vision has continually developed proprietary AI algorithms that power its products and solutions. Hanwha Vision’s ongoing collaboration with institutions like the Gwangju Institute of Science and Technology (GIST) strengthens the synergy between academia and industry, contributing to the further advancement of AI technologies.

Hanwha Vision is the leader in global video surveillance with the world's best optical design / manufacturing technology and image processing technology focusing on video surveillance business for 30 years since 1990.