Philips and LG Uplus Also Found Startup Problem Solvers (Comprehensive)

Philips 'AI-Based Patient Condition Analysis', LG Uplus 'Mobile Baseball Broadcast Service Vividness Maximization' Project Finals Results

On the 20th, Kang Sung-cheon, Vice Minister of the Ministry of SMEs and Startups, is delivering a greeting at the Daestar Solver Platform Demo Day.

[Asia Economy Reporter Kim Cheol-hyun] Startup problem solvers are emerging one after another to address the concerns of large corporations. Philips introduced its 'AI-based patient condition analysis,' and LG Uplus presented its challenge to 'maximize the vividness of mobile baseball broadcasting services,' meeting startups to solve these tasks together.

The Ministry of SMEs and Startups (Minister Park Young-sun, hereafter MSS) announced on the 20th that it held the first phase of the Large Corporation-Startup Problem Solver Platform contest for Philips Korea's 'AI-based patient condition analysis' and LG Uplus's 'maximizing vividness of mobile baseball broadcasting services,' selecting Medipixel, Meju, Vuno, Lab2Eye, and PixelScope respectively.

The 'Large Corporation-Startup Problem Solver Platform' is a new win-win cooperation policy that connects large corporations and startups to solve each other's problems and complement their different capabilities. In the first contest, 208 startups participated in challenges from nine large corporations. From July 30, over four months, various startup ideas were verified, and the feasibility of solving the challenges was reviewed. Forty-three startups were selected for the final round, and since the 16th, finals for eight challenges have been ongoing.

◆ Startups to solve Philips' 'AI-based patient condition analysis' challenge = In this final round, startups were selected to solve Philips Korea's challenge of 'AI-based patient condition analysis' for non-contact monitoring of infectious disease patients and hospitalized patients. Philips prioritized applicability and commercialization potential in medical settings to enable rapid adoption, given the urgent need to protect medical staff amid the COVID-19 pandemic. The six startups participating in the finals demonstrated various AI-based technologies, with medical professionals from diagnostic laboratory medicine and infectious diseases involved in the evaluation.

The final selectee Medipixel proposed a technology that shortens analysis time by applying deep learning-based automatic segmentation technology to angiography result analysis, which was previously done manually. This technology attracted attention from medical evaluators as it can assist clinical decision support, such as lesion diagnosis and optimal stent tool recommendation.

Meju showcased technology that uses a body-attached ultra-small electrocardiograph (HiCardi) to measure nine types of biometric data in real-time, including heart rate, respiration rate, and activity level, enabling medical staff to monitor multiple patients remotely simultaneously. Notably, Meju's self-developed HiCardi has completed domestic regulatory approval and is currently undergoing nine clinical trials, making commercialization feasible in the near future, earning high marks.

Vuno presented AI software that analyzes patient data such as pulse, blood pressure, respiration, and body temperature using the latest deep learning technique, recurrent neural networks, to predict cardiac arrest and heart disease early. Compared to the existing 'early warning score' technology, it showed 19% higher prediction accuracy and a 63% reduction in false alarm rate, highlighting its distinctiveness. Kang Sung-chun, Deputy Minister of MSS, said, "Digital healthcare, whose importance has grown due to COVID-19, is a future new industry. Since it is a field where medical information services and information communication technology converge, combining Philips' network with medical institutions and medical device manufacturing experience with startups' AI technology through the Large Corporation-Startup Problem Solver Platform is expected to create new growth engines."

◆ LG Uplus' 'Maximizing vividness of mobile baseball broadcasting service' challenge = The final round seeking solutions for LG Uplus's 'maximizing realism of mobile baseball broadcasting dedicated app service' included former LG player Kim Yong-su as a judge, who carefully examined the realism of the baseball game and the validity of pitch prediction. Three companies participated, showcasing technologies such as real-time extraction of pitch and hit data using high-speed cameras and video processing visualized as broadcast content, AI analysis of pitcher posture for pitch prediction, and high-resolution restoration technology to vividly enlarge every corner of the stadium.

Among them, Lab2Eye combined video and player motion information collected at the game site to visualize strike zones in live broadcast videos and demonstrated a method to predict batter pitch types through AI analysis trained on pitcher posture and game situations from multiple baseball game videos. While the face-to-face evaluation distinguished two pitch types (fastball and breaking ball), the final round showed an algorithm predicting nine pitch types, earning high scores for future development potential.

PixelScope provided a service that visualizes real-time analysis results of pitch and hit trajectories from baseball game videos transmitted from cameras installed in the stadium and offers split-screen views enlarging players' baserunning and defense through virtual cameras. It was evaluated as excellent in AI video analysis and real-time result derivation capabilities.

The selected startups will receive up to 2.5 billion KRW in government support, including commercialization funding (100 million KRW), technology development (up to 400 million KRW), and technology special guarantee (up to 2 billion KRW). They will also gain opportunities to jointly promote projects with large corporations and expand globally using overseas networks.