Lingyi Zhihui Launches Industry's First Group Standard for AI-based Fundus Image Annotation to Advance Ophthalmic AI Standardization

On September 21, 2019, at the Second National Conference on Artificial Intelligence in Ophthalmology, guided by the China Medical Education Association and supported by Baidu Brain technology, the AI healthcare brand “Lingyi Zhihui” jointly launched the industry’s first group standard, “Annotation Standards for AI-Based Color Fundus Photography,” led by the Zhongshan Ophthalmic Center of Sun Yat-sen University. The “Annotation Standards for AI-Based Color Fundus Photography” will extensively engage experts from industry, academia, research, and clinical practice to ensure data quality at the source and promote the standardized development of the AI ophthalmology industry.

At the event, Professor Zhang Xiulan, Council Member of the Asia-Pacific Glaucoma Society and Director of the Clinical Research Center at Zhongshan Ophthalmic Center, Sun Yat-sen University, was invited to deliver a speech as the representative of the launch ceremony. Experts in attendance included Ms. Xu Shan, Senior Business Manager at the China Academy of Information and Communications Technology (CAICT) and Vice Chair of the WHO-ITU Focus Group on Artificial Intelligence for Health; Professor Zhang Shaochong, Director of the Special Needs Medical Center at Zhongshan Ophthalmic Center, Sun Yat-sen University; Mr. Yang Weihua, Deputy Chairman and Secretary-General of the Intelligent Medicine Professional Committee under the Chinese Medical Education Association; Professor Liu Jiang from the Department of Computer Science and Engineering at Southern University of Science and Technology; and Professor Tan Mingkui, Director of the Computing Center at South China University of Technology. Together with Lingyi Zhihui, they witnessed this significant moment.

“AI Fundus Color Photography Annotation” Group Standard Launch Ceremony Expert Group Photo

It is reported that the group standard titled “AI Annotation of Color Fundus Photographs” is an enhanced and upgraded version of the “Enterprise Standard” released by Lingyi Zhihui on September 5 this year. This Enterprise Standard was jointly developed under the leadership of Lingyi Zhihui and the Zhongshan Ophthalmic Center of Sun Yat-sen University. Dozens of senior ophthalmologists, data annotation physicians, and AI R&D professionals participated in its formulation, striving to balance clinical professionalism with the feasibility of algorithm training. Ophthalmology experts and AI researchers conducted multiple rounds of in-depth discussions, focusing on the goal of “aligning with clinical practice, being suitable for algorithm training, and enabling quantifiable and accurate execution,” while addressing practical issues encountered during annotation and algorithm training processes, thereby producing an initial draft. The draft was subsequently refined through meticulous review of details, followed by multiple rounds of audit and proofreading by the expert team, ultimately resulting in a comprehensive and precisely detailed final version. This final version was released as an Enterprise Standard at the “24th National Ophthalmology Academic Conference of the Chinese Medical Association.”

Group standards represent an upgraded version of enterprise standards. This group standard not only defines the annotation and referral rules for diabetic retinopathy, glaucoma, macular lesions, and high-risk pathological myopia, but also provides detailed explanations on handling inconsistent annotations and standardizing them for artificial intelligence algorithm training. Furthermore, it extends and redefines the annotation of “age-related macular degeneration” by expanding its scope to “macular lesions,” thereby better aligning with referral objectives in practical applications.

Following the launch ceremony, Lingyi Zhihui invited experts and guests from regulatory bodies, research institutions, hospitals, and enterprises to participate in a seminar titled “How the Group Standard for ‘AI Color Fundus Photography Annotation’ Facilitates the Standardized Development of the Ophthalmic AI Industry.” The attending experts and guests highly endorsed Lingyi Zhihui’s initiative to lead the development of data annotation standards. They provided multi-dimensional interpretations of the establishment and promotion of the Group Standard for “AI Color Fundus Photography Annotation,” and, drawing on their respective professional experiences, shared their expectations and recommendations for the standardization of the ophthalmic AI sector and the broader medical AI industry.

Group Photo of Experts at the Seminar on “How the Group Standard ‘AI Annotation of Color Fundus Photography’ Facilitates the Standardized Development of the Ophthalmic AI Industry”

Reflecting on the development of the standard, Professor Zhang Xiulan, Director of the Clinical Research Center at Zhongshan Ophthalmic Center, shared her profound insights: “The creation of this standard has been far from easy. It truly enables mutual understanding and effective dialogue between clinicians and AI practitioners. I often say that standards must precede industry development. Without standards, how can we obtain high-quality data? And without high-quality data, how can we develop AI algorithms that are genuinely practical and effective? Zhongshan Ophthalmic Center and Lingyi Zhihui are committed to sharing our achievements, hoping that more teams will adopt them and that more experts will join this endeavor.”

Discussing the significance of enhancing data quality through standardization, Wang Bing, Product Architect at Lingyi Zhihui, shared her profound insights. She pointed out, “AI algorithms are largely data-driven; without high-quality data, it is difficult to build a robust product, which is a common challenge across the industry. AI in healthcare holds profound significance, attracting numerous companies to this competitive landscape. However, amidst this race, no one has yet paused to effectively address these industry-wide issues. As a participant in this sector, Lingyi Zhihui bears the responsibility to dedicate effort and time to resolving them by establishing standards for annotation and quality control of AI-based color fundus photography. This is particularly crucial for the overall development of the industry.”

Chairman Xu Shan, representing the policy and regulatory perspective, interpreted the significance of the standard as follows: “Standardization is one of the key means to promote the sound development of the industry. From a standardization standpoint, the AI healthcare sector is still relatively new. Therefore, tackling the challenging issue of data quality holds substantial significance for the launch of this standard. Lingyi Zhihui’s pioneering exploration in ophthalmic data quality control will provide valuable experience for other specialized fields of medical artificial intelligence. We are also committed to providing continuous support for the internationalization of this standard.”

Professor Yang Weihua of the Intelligent Medicine Professional Committee under the China Medical Education Association has participated in the development of multiple medical databases. He offered a positive assessment of the “Standard for Annotation of AI-Based Color Fundus Photography.” “In my view, this standard is highly necessary. It serves as a common professional language for communication among all parties collaborating on AI research. With this bridge in place, stakeholders can achieve a consistent understanding at the outset of AI research initiatives. The upgraded group standard will help accelerate ophthalmic AI research and may facilitate the advancement of research paradigms in this field.” As a representative of the organization supporting this group standard, Professor Yang’s endorsement fully reflects the strong support provided by the Intelligent Ophthalmology Group, the Intelligent Medicine Professional Committee, and the China Medical Education Association.

Professor Liu Jiang of Southern University of Science and Technology shared his perspective from the standpoint of artificial intelligence research: “I am very pleased to see Lingyi Zhihui launch this standard. It is no small feat for such a large company to start with such detailed and seemingly minor aspects. I believe that this standard will certainly drive the industry forward.”

Lingyi Zhihui is a smart healthcare brand powered by Baidu Brain’s AI technology. With the vision of “empowering primary healthcare through evidence-based AI,” it is committed to serving all scenarios both within and outside hospitals, helping industry partners such as hospitals, HIS vendors, electronic medical record (EMR) vendors, and health management operators improve quality and efficiency. The release of this standard will further solidify Lingyi Zhihui’s position as an industry leader.

In fact, the promotion of the group standard “AI Annotation of Color Fundus Photographs” is just one of the efforts made by Lingyi Zhihui to promote the standardization of the entire industry. At the “2019 Global Artificial Intelligence Health Summit” held in August this year, the World Health Organization (WHO) and the International Telecommunication Union (ITU) of the United Nations officially established the Domestic Promotion Group for the Focus Group on AI for Health. The “Focus Group,” established last year, is dedicated to preliminary research on international standards for artificial intelligence in healthcare, with a focus on data formats, standard datasets, and algorithm evaluation and validation in health and medical AI. As a deputy leader of the domestic counterpart group of the Focus Group, Lingyi Zhihui has participated actively and strives to promote deep collaboration among various Chinese stakeholders in jointly developing international standards for medical artificial intelligence.

Earlier this year, Lingyi Zhihui, as the lead organization, submitted two standard proposals to the AI for Health Focus Group, covering fundus imaging and clinical decision support systems, which received high recognition from international experts and peers within the Focus Group. In May of this year, at the “Third Global Summit on Artificial Intelligence for the Benefit of Humanity” held by the United Nations, Lingyi Zhihui participated in the Focus Group’s discussions at the summit, further advancing standards development in collaboration with international counterparts.

Lingyi Zhihui firmly believes that only by being down-to-earth and treating every link and element of “medical care” with meticulous professionalism can AI in healthcare truly benefit the public. As a relevant representative from Lingyi Zhihui stated, “This standard serves more as an initial effort to spark further discussion and development. We will continue to promote and assist in the establishment of multimodal data standards for ophthalmology, potentially extending to related standards in other disease areas. Our goal is to create products tailored to a wide range of primary-care application scenarios, empowering the entire industry toward standardized and progressive development with a professional and rigorous approach.”

The path to standardization is not easy. The launch of the first group standard, “Annotation Standards for AI-Based Color Fundus Photography,” will usher in a new chapter for the ophthalmic AI industry. It is believed that more enterprises, hospitals, and researchers will join this group standard in the future to jointly build a regulated and standardized ophthalmic AI industry, thereby facilitating its sustained and robust development.