This page was machine-translated and may differ from the original. View original
KETI wins first place in IEEE JPEG Trust Challenge…invisible watermarking technology reaches world-class level
Competition with 12 participating countries: winning first place ahead of U.S. company Trufo, achieving results in self-supervised learning-based technology
The Korea Electronics Technology Institute (KETI) has won first place in a challenge held under an international academic conference with its invisible watermarking technology. As demand for digital content authenticity verification grows in response to the spread of generative AI, this case of a domestic research institute confirming its international competitiveness in the related technology field draws attention.
KETI announced on the 21st that its Composite Intelligence Research Center research team won first place at the JPEG Trust Watermarking Benchmark Challenge at the 2026 IEEE International Conference on Image Processing (ICIP 2026) held in Tampere, Finland on the 16th (local time).
The competition was attended by companies and institutions from 12 countries including the United States, United Kingdom, Germany, and France. In the final round, the research team competed against Trufo, a U.S. startup specializing in content authenticity, and won first place.
The JPEG Trust Challenge is held to evaluate the performance of invisible watermarking technology required in the generative AI era and promote technological advancement.
Invisible watermarking is a security technology that embeds information in digital content such as images and video in a manner that is not visually discernible to the human eye.
The competition comprehensively evaluated whether watermarks could be stably detected under various conditions including compression, cropping, geometric transformation, signal processing, and transformation using generative AI, based on international standards (ISO/IEC 21,617).
The KETI Composite Intelligence Research Center research team submitted technology that finely distributes and detects watermarks in key areas of images based on a self-supervised learning model.
Cho Chung-sang, principal researcher who led technology development, stated, "We will continuously advance our technological capabilities to respond to increasingly complex content transformation techniques."
Lee Young-han, director of the Composite Intelligence Research Center, said, "We will contribute to building a reliable media ecosystem by spreading the related technology to the industry and enhancing its sophistication."
This research was conducted through national research and development projects supported by the Institute for Information & Communications Technology Planning & Evaluation (IITP).
To request a correction, reply or follow-up report on this article, see how to file a request. Previously published statements are collected in corrections & replies.













