Former Baidu Executive's AI-Powered Handheld Retinal Camera Gains Traction in Over 200 Grassroots Clinics Following Tens of Millions in Funding

By Luo Mei, Nie Guanghong

In the field of fundus screening, Beijing Zhizhen Internet Technology Co., Ltd. has launched the world’s first handheld, portable “Intelligent Fundus Screening Camera,” designed specifically for primary care physicians to enable a simpler, more portable, and intelligent imaging process.

According to founder Dai Liming, this smart device is about to be deployed in hundreds of primary healthcare institutions. The plan for 2019 is to empower thousands of primary healthcare service providers, facilitating the implementation of eye health management services at the grassroots level.

Dai Liming spent a decade at Baidu, where he served as an executive in the company’s commercial division. The pre-A financing round was also led by Baidu. He stated that this collaboration with his former employer marked a technological turning point for Zhizhen. This makes Zhizhen one of the few healthcare companies in the AI sector to receive early-stage investment from Baidu.

Meanwhile, last year, the company completed a tens-of-millions RMB Pre-A financing round, jointly led by Baidu and Shengjing Jiacheng, with participation from Meihua Angels, Huayao Capital, and Huajian Capital, while Evolution Capital served as the exclusive financial advisor.

So, how was this AI fundus screening camera successfully developed? How is the imaging performed? What are its key highlights? VCBeat (WeChat ID: vcbeat) conducted an exclusive interview with Dai Liming, CEO and co-founder of Zhizhen.

For a long time, “poor portability,” “difficulty in image acquisition,” and “challenges in image interpretation” have been the three major pain points of traditional fundus screening in primary care settings. “It would be ideal to have a ‘point-and-shoot camera’-style screening device that is easy to operate, allowing clinicians to determine whether there are any abnormalities simply by taking a picture of the eye.” This sentiment is widely shared among primary care physicians.

Dai Liming introduced that the intelligent screening fundus camera is specifically designed for fundus screening by primary care physicians. Integrating intelligent fundus imaging and disease screening, it was jointly developed by a team of 100 R&D personnel over two years to meet the core needs of primary care physicians in screening scenarios. It features three core technologies.

First, intelligent autofocus technology. He stated that the biggest challenge in developing smart fundus screening cameras was how to intelligently control the optical path. They aimed to direct light through the pupil at a 45-degree angle to observe the human fundus, making the selection of the light source a critical issue.

According to Dai Liming, performing mydriasis at the primary care level carries significant risks; therefore, a non-mydriatic approach is optimal, enabling light to be projected into the fundus without pupil dilation. In collaboration with their partners, they have optimized the optical path by adopting a near-infrared light mode. Meanwhile, intelligent autofocus technology adjusts clarity in real time, ensuring that the smart camera directs near-infrared light into the fundus via the shortest possible time and optimal pathway, thereby yielding high-resolution images.

Second is intelligent tracking technology. Because fundus photography must meet diagnostic standards, to ensure stable acquisition of key areas in the fundus, the camera performs semantic recognition and real-time tracking during the imaging process, ensuring that all critical information in the captured fundus images is acquired. This is a major challenge for many primary care physicians when learning to use traditional equipment. With intelligent tracking technology, the camera can assist primary care physicians in accurately locking onto the target position and capturing images.

Third is intelligent exposure technology. After the device is aligned with the eye, a single button press initiates the acquisition process, which takes only a few hundred milliseconds. However, numerous issues can arise within this brief window, such as physician hand tremor or patient eye movement, resulting in substandard images that are unsuitable for diagnosis. Therefore, the camera performs real-time image quality assessment and automatically captures the image only when the quality score exceeds 90 points, thereby ensuring a high pass rate for the final fundus images.

Finally, the smart fundus camera is equipped with 4G and Wi-Fi connectivity, enabling real-time data transmission to the cloud. The data is analyzed by an AI-assisted algorithm engine and subsequently reviewed by physicians, allowing for the generation of medical-grade screening reports within five minutes. This not only enhances the patient experience but also significantly improves healthcare efficiency for primary care physicians, facilitating follow-up services.

This device has also brought about changes to Zhizhen’s business model. Dai Liming stated that, based on the “Intelligent Fundus Screening Camera,” Zhizhen Interconnect locked in two strategies at the beginning of this year, with the first strategy focusing on how to achieve commercialization. Having previously served as the head of Baidu’s commercial division, Dai Liming possesses a clear understanding of how to construct a sound business logic.

Full commercialization is characterized by three key elements. The first is demand, which is fundamental; human demand for health-related services is substantial. The second is whether a company can demonstrate its commercial value. Simply put, this means the company provides services that meet consumer needs, consumers are willing to engage in equivalent exchange for these services, and health insurance providers are also willing to pay for them. The third is whether the commercial aspects of this process can be converted into corporate profits.

Among these factors, profitability is a key issue; if the application scenarios for the equipment cannot be expanded, it will be impossible to generate profits. Therefore, Zhizhen started with the second point by partnering with hardware companies to develop its own hardware.

However, the high costs associated with hardware bundling pose a significant challenge. Given these substantial expenses, Zhizhen still needs to establish a marketing team. Therefore, to ensure future profitability, cost pressures must be intensified. In this process of cost reduction, the first priority is to lower equipment prices, and the second is to enhance overall operational efficiency through empowerment at the grassroots level. Only in this way can the equipment be more effectively promoted, creating a win-win situation for both the company and its customers under this system. Currently, Zhizhen’s strategy is to deploy this equipment free of charge to primary healthcare institutions and then share the resulting revenues with them.

Regarding the specific application model of the “Intelligent Screening Fundus Camera,” Dai Liming stated that collaborations with primary healthcare institutions are currently primarily in the model-building phase. At present, more than ten primary healthcare institutions have begun using this device, and the number of contracted medical institutions has reached approximately 200. It is expected that over 100 medical institutions will have the device fully operational before the Spring Festival. The plan for 2019 is to reach 1,000 institutions.

“Light is ordinary for healthy people, but a luxury for patients.” This is why Dai Liming chose ophthalmology as the field for his entrepreneurial venture.

It was May 2016. On that day, Dai Liming, serving as a volunteer, accompanied a team to Suining County in the Xiangxi region of Hunan Province to conduct a free clinic. During the event, an eleven-year-old girl arrived with her eyes closed, supported by her father. The medical staff examined her using traditional fundus screening equipment. Experts from Beijing provided a remote diagnosis, concluding that she had no light perception for life, meaning complete blindness.

This case shocked Dai Liming. After repeatedly confirming with experts in Beijing, the results remained consistent with the initial findings. The experts explained that the young girl had a family history of congenital glaucoma; however, due to the lack of timely detection, her condition deteriorated, leading to blindness. Had this patient been diagnosed two years earlier, she would not have ended up in her current state.

The girl listened calmly to the specialist’s diagnosis, neither crying nor making a fuss, and walked out with her father’s support.

At this point, Dai Liming’s eyes welled up with tears. Deeply shaken by the incident, he began to reflect: during his ten-year tenure at Baidu, where he had access to vast technological resources and expertise, could he make meaningful contributions to ordinary people?

Due to the large crowd, the free clinic, which was scheduled to end at 12:30 p.m., was extended until 1:30 p.m. After the clinic concluded, as they stepped outside, they found that local patients were still waiting there. The elderly joined their palms together in a gesture of gratitude and watched them depart.

It was in that very instant that Dai Liming saw the value and significance of doing this.

“Smart Screening, Safeguarding Vision”—these eight words feature prominently on the homepage of Zhizhen’s official website. Dai Liming told us that he coined this phrase on the return journey after a free clinic day, gazing out at the drizzling rain through the car window. These eight words have since become Zhizhen’s steadfast professional creed.

“If your products and services can help these patients preserve their vision, you are truly benefiting them,” said Dai Liming.

Previously, Dai Liming volunteered with medical equipment to provide free clinical consultations alongside grassroots doctors in Xuyong County, Sichuan Province, a nationally designated poverty-stricken county. During the screening process, they encountered an elderly patient who required eye surgery. Although the procedure cost only 2,000 yuan, Dai and his team spent three days trying to persuade the elder to undergo the operation, but he firmly refused. The elderly man explained that despite his family’s financial hardships, he could afford the 2,000 yuan; however, he was unwilling to spend it on himself in his twilight years and insisted on saving the money for his granddaughter’s tuition fees.

Dai Liming said that in that instant, he wondered why treating this disease cost 2,000 yuan instead of 200 yuan. It is precisely because these conditions cannot be detected early in their initial stages, leading to deterioration in later stages and necessitating more complex and expensive treatments.

He discussed the challenges of working at the grassroots level with primary care physicians. One doctor’s words deeply moved him: “We may be just ordinary grassroots doctors, but we are busy every day and eager to learn advanced techniques and methods. Why do we work so tirelessly? Because we believe that saving one person is equivalent to saving an entire family. If no one steps forward to help them, then no one will. It is precisely because professionals like us are willing to join this field, each contributing our share and lending a helping hand, that we can make a difference.”

These experiences are key drivers of Zhizhen’s intrinsic motivation. Following this, Dai Liming established Zhizhen’s goal of “serving grassroots healthcare.” Investing capital into intelligent products to enhance their performance and metrics, thereby simplifying the disease detection process and improving imaging accuracy, would carry greater significance. This aligns with the values upheld by Zhizhen.

To enable universal eye health screening services for people with diabetes in more grassroots areas, Zhizhen specially donated 200 intelligent fundus screening cameras to the China Association of Health Care for the Elderly at the press conference, to assist endocrinology departments in county-level regions in carrying out efficient and high-quality diabetic retinopathy screening work.

Professor Liu Yuanli, Dean of the School of Public Health at Peking Union Medical College and President of the China Health Care Association for the Elderly, stated that endocrinology departments are currently the initial point of contact for diabetic retinopathy screening in many primary care hospitals; however, the proportion of these departments equipped with screening capabilities remains very low. The introduction of intelligent fundus cameras into primary care endocrinology departments addresses both image acquisition and diagnostic challenges. It is hoped that these intelligent devices will benefit a broader population of patients in primary care settings and contribute to the accumulation of more extensive experience in diabetes prevention and control management across China.

Dai Liming stated that the penetration rate of intelligent devices will increase significantly, enabling eye health screening at every primary care hospital, family practice, and even in pharmacies and optical shops. In the future, these devices may even become available for household use, allowing individuals to perform self-screenings at home, with reports uploaded to the cloud for AI analysis and remote interpretation. At that stage, eye health screening and management will become more inclusive and accessible to everyone, thereby enhancing its social significance and value.

This device empowers the data entry point, enabling even primary care physicians to use it simply and efficiently. This constitutes the product’s core value proposition and competitive advantage.

As data accumulates to higher levels, all knowledge systems in the field of bioinformatics must converge on a single focal point: scientific discovery. In this context, such scientific discovery refers to the correlation between fundus images and clinical outcomes. Dai Liming predicts that once scientific breakthroughs are unlocked through fundus imaging, they will inevitably give rise to a great enterprise. He also anticipates that around 2020, a wide range of related application scenarios will emerge, including the adoption of these devices in home settings.

In the future, Zhizhen will extend its AI capabilities to a broader range of diseases and plans to conduct big data analytics on its existing fundus screening data. Dai Liming informed us that Zhizhen’s database is already a global leader in scale, though this currently covers only adults. This September, Zhizhen will establish a leading domestic neonatal fundus imaging database. Neonatal fundus screening can detect various ocular diseases, most of which are curable through early screening, intervention, and treatment.

Concurrently, the “Intelligent Fundus Screening Camera” is undergoing global medical certification; if approved, this device will open up international markets for Zhizhen.