Xiao-I's Virtual Try-on Technology (OOTDiffusion) Achieves Significant Breakthroughs, Attracting Industry-wide Attention and Sparking Discussions

News provided by

Apr 23, 2024, 09:49 ET

OOTDiffusion Product is Expected to be Launched in May

SHANGHAI, April 23, 2024 /PRNewswire/ -- Xiao-I Corporation (NASDAQ: AIXI) ("Xiao-I" or the "Company"), a leading company in the field of artificial intelligence, is pleased to announce its recent significant progress in its latest product line, virtual Try-on technology (OOTDiffusion). The OOTDiffusion algorithm, independently developed by Xiao-I, was released on arXiv, an open-access repository of electronic preprints and postprints, in March 2024 (Access Paper: https://arxiv.org/abs/2403.01779).

OOTDiffusion is a new network architecture for realistic and controllable virtual try-on (VTON). Xiao-I utilizes a pre-trained latent diffusion model and designs an equipped UNet to learn the detailed features of clothing. Through the outfit fusion method proposed in the self-attention layers of the denoising UNet, precise alignment of clothing features with the target human body can be achieved without redundant deformation processes. To enhance controllability, Xiao-I introduces outfit dropping, which adjusts the intensity of clothing features during the training process without the need for a classifier. Comprehensive experiments conducted by Xiao-I on the high-quality VTON datasets demonstrate that OOTDiffusion can efficiently generate high-quality try-on results applicable to any human body and clothing image. It outperforms other VTON methods in terms of realism and controllability, marking a remarkable breakthrough in the field of virtual try-on.

OOTDiffusion has generated widespread attention and interest in the industry. On February 23, 2024, Xiao-I has released the algorithm on GitHub, an online software source code hosting platform (project homepage: https://github.com/levihsu/OOTDiffusion). As of now, the project has received approximately 4.4k stars and 620+ forks, demonstrating the high recognition and attention from the open-source community. Furthermore, the algorithm topped the "Hugging Face Paper 3.5-Day Leaderboard", a machine learning application tools development company in the United States, and sparked extensive discussions on X, formerly known as Twitter.

Xiao-I is dedicated to making changes in everyone's life with AI, driven by the increasing demand for virtual try-on technology in the market. The OOTDiffusion product is expected to be launched on Xiao-I's official website in May. This product will provide users with a brand-new dressing experience, allowing them to try different clothing styles in a virtual world using intelligent algorithms and virtual reality technology. Whether it is a source of fashion inspiration or adding fun to the shopping experience, OOTDiffusion will become an essential tool in users' lives. Xiao-I will continue to drive the development of AI technology, bringing more innovation and surprises to users. Please stay tuned for the release of the OOTDiffusion products in May by Xiao-I and enjoy the new experience brought by virtual try-on.

About Xiao-I Corporation

Xiao-I Corporation is a leading cognitive intelligence enterprise in China that offers a diverse range of business solutions and services in artificial intelligence, covering natural language processing, voice and image recognition, machine learning and affective computing. Since its inception in 2001, the Company has developed an extensive portfolio of cognitive intelligence technologies that are highly suitable and have been applied to a wide variety of business cases. Xiao-I powers its cognitive intelligence products and services with its cutting-edge, proprietary AI technologies to enable and promote industrial digitization, intelligent upgrading, and transformation. For more information, please visit: www.xiaoi.com.

Forward-Looking Statements

This press release contains forward-looking statements as defined by the Private Securities Litigation Reform Act of 1995. Forward-looking statements include statements concerning plans, objectives, goals, strategies, future events or performance, and underlying assumptions and other statements that are other than statements of historical facts. When the Company uses words such as "may," "will," "intend," "should," "believe," "expect," "anticipate," "project," "estimate" or similar expressions that do not relate solely to historical matters, it is making forward-looking statements. Forward-looking statements are not guarantees of future performance and involve risks and uncertainties that may cause the actual results to differ materially from the Company's expectations discussed in the forward-looking statements. These statements are subject to uncertainties and risks including, but not limited to, the following: the Company's ability to achieve its goals and strategies, the Company's future business development and plans for future business development, including its financial conditions and results of operations, product and service demand and acceptance, reputation and brand, the impact of competition and pricing, changes in technology, government regulations, fluctuations in general economic and business conditions in China, and assumptions underlying or related to any of the foregoing and other risks contained in reports filed by the Company with the U.S. Securities and Exchange Commission ("SEC"). For these reasons, among others, investors are cautioned not to place undue reliance upon any forward-looking statements in this press release. Additional factors are discussed in the Company's filings with the SEC, including under the section entitled "Risk Factors" in its annual report on Form 20-F, as amended, filed with the SEC on August 10, 2023, as well as its current reports on Form 6-K and other filings, all of which are available for review at www.sec.gov. The Company undertakes no obligation to publicly revise these forward-looking statements to reflect events or circumstances that arise after the date hereof.

SOURCE Xiao-I Corporation