4by4, Image Quality Enhancement AI Research Paper Published in SCI-Level Journal

Leading In-House AI Research Lab Develops Advanced Image Quality Enhancement AI Solution
Proposes Improved Methods for Image Quality Enhancement

Content AI solution provider ForBy4 announced on the 22nd that a research paper on a new image quality enhancement AI model developed by its in-house AI research lab, Pixell Lab, has been published in the global scientific journal Scientific Reports.

Scientific Reports is a sister journal of Nature, one of the world's top three scientific journals. It is classified as an SCI (Science Citation Index) level journal, the highest grade among prestigious journals selected by the U.S. Institute for Scientific Information (ISI) based on strict criteria.

The study titled "Research on Multi-Color Space Networks for Image Enhancement Using Complex Comprehensive Data" introduces a newly devised method to address the issue that images taken in overly dark or bright environments can affect the 'color' of the resulting image.

The new model proposed by ForBy4 AI Lab overcomes the limitations of AI models that simply recognize the RGB color space and repeatedly apply image enhancement processing without considering different lighting conditions. It first calculates and processes intermediate data of average brightness values that fully capture color information. To achieve this, it suggests sequentially applying two separate networks operating in different color spaces: a Transitional Network that calculates the average brightness value, followed by a Base Network for image quality enhancement.

Kim Hyun-chul, head of ForBy4 AI Lab who led the paper writing and validation process, said, "Objective survey results confirmed that the outputs improved using the multi-color space network showed overwhelmingly superior performance across various datasets compared to those improved by single color space networks."

He added, "The newly proposed network model not only produces visually high-quality effects but also proves to be very effective in generating more realistic video images, that is, lifelike images similar to actual scenes."

Lim Jeong-hyun, vice president of ForBy4 in charge of the Pixell business division, explained, "The AI model introduced in the paper has already been applied to our AI solution Pixell and is receiving favorable evaluations in the market with performance surpassing other models." He emphasized, "It can be immediately applied across industries, such as effectively enhancing CCTV footage taken in dark environments to better identify subjects and assist in criminal investigations, demonstrating its vast potential and utility."

He also expressed his ambition, stating, "Along with this published paper, we will actively promote that ForBy4's technological capabilities meet global standards through various means, such as the 'sharpness enhancement method patent' currently filed under PCT (USA) international application."

This content was produced with the assistance of AI translation services.