Deepnoid (CEO Choi Woo-sik), Korea's first-generation medical artificial intelligence (AI) company, announced that it’s going to begin constructing a telemedicine AI system by combining large language model (LLM) model technology with its current vision AI-centered medical solution.
Jae-ik Cho, Chief Operating Officer of Deepnoid, said, “We’ll transcend video-based reading assistance solutions and launch latest products using Large Language Model (LMM) technology. If we gather all of the technology and know-how, telemedicine AI won’t be distant.” “I’m watching,” he said on the 18th.
Deepnoid can also be showing remarkable ends in the LLM field, rating first with a mean of 61 points within the 'Open Ko-LLM Leaderboard', an evaluation index for Korean language models, last week.
COO Cho said, “Along with commercializing LLM-based services, we’re also pursuing AI solutions using non-image data. We’re continuing research in collaboration with excellent domestic university hospitals and medical device corporations, so we are going to give you the option to launch it soon.” .
Once LLM is introduced, quite a lot of comprehensive medical services beyond existing image evaluation will develop into possible. “Ultimately, medical AI can have no selection but to maneuver towards a ‘distant reading specialist hospital’ or a ‘telemedicine platform,’” he explained.
He cited the next as vital conditions for establishing a telemedicine system: ▲ accumulating data through cooperation with actual medical institutions ▲ having various solutions for every area and symptom, and explained, “Deepnoid satisfies each conditions.”
First, Deepnoid is developing an AI solution that supports disease diagnosis, prediction, and treatment based on 17 'Deep:AI' products and a couple of 'Deep:PAX' products which have obtained approval from the Ministry of Food and Drug Safety. They said they plan to offer a 'total solution' that features medical image reading services in other fields in the long run.
In addition they identified that since they entered the medical market relatively early, the related data and cooperation system are solid.
Since 2020, we have now established Huain Imaging Clinic, a hospital specializing in AI image reading, and have been providing distant medical treatment and image reading services on site. As well as, we have now established a DeepPax medical system on the Health Insurance Review and Assessment Service (HIRA) and are conducting joint research and construction of AI solutions with major domestic general hospitals.
Meanwhile, Deepnoid, founded in 2008, began with Web of Things (IoT) technology, went on to industry-specific AI akin to 'equipment failure prediction solutions', and is currently specializing in medical AI, its fundamental business field.
He said that he had been preparing for the business because the time when even the concept of medical AI was rare. COO Cho said, “The more I learned about AI, the more I believed there could be useful fields. Specifically, if applied to the medical field, I believed it could solve many problems, from image evaluation, diagnosis, and treatment plan development.”
Consequently, starting with Korea's first spinal disease AI diagnosis assistance solution 'Deep: Spine', various products akin to 'Deep: Neuro' to help in diagnosing cerebral aneurysms, 'Deep: Chest' to help in lung disease diagnosis, and 'Deep: Lung' to help in pulmonary nodule diagnosis. An answer was presented.
He said that he’s currently specializing in developing overseas markets for his business.
We’re discussing business with national medical institutions in Uzbekistan, the Philippines, and Indonesia, and are actively conducting sales by installing demo solutions. As well as, the corporate sought to enter the Middle East market through 'Arab Health 2024', the most important medical device exhibition within the Middle East.
Through this, we’re preparing an MOU for research cooperation, publishing papers in international academic journals, and obtaining FDA and CE certification.
Amongst them, Southeast Asia is an emerging market following america and Europe, and is the goal of current focus. Based on the Korea Medical Device Safety Information Institute, the medical device market in major Southeast Asian countries akin to Thailand, Malaysia, and Vietnam is anticipated to achieve $2 billion (roughly 2.6 trillion won) in 2024 alone.
In Korea, Deep: Neuro, which was chosen as an progressive medical device last yr, has been introduced to 7 hospitals. As well as, the establishment of a nationwide sales organization was accomplished.
Although the domestic medical AI market is in a comparatively early stage, it is anticipated to have high growth potential. Based on the Korea Health Industry Development Institute, the domestic medical AI market is anticipated to grow to about $12.2 billion (about 16.25 trillion won) in 2027.
COO Jae-ik Cho said, “If we mix all of the technology and know-how, distant medical AI appears to be not distant. We’ll construct an answer in order that any medical staff can use AI as easily as Excel.”
He also said, “We’ll give you the option to see tangible results starting within the second half of this yr,” and added, “We predict that there will likely be meaningful sales within the medical AI field this yr.”
Reporter Jang Se-min semim99@aitimes.com