Superb AI supplies smartphone-based caries diagnosis AI to Ajou University Dental Hospital

Superb AI (CEO Hyunsoo Kim), a vision AI all-in-one solution company, announced on the 13th that it has supplied an AI model for diagnosing caries using smartphones to the research team of Professor Heekyung Kim at Ajou University Dental Hospital.

The newly developed caries diagnosis AI model detects caries early using smartphone images to assist in the diagnosis of specialists. It was developed to enable easy detection of caries even in areas where it is difficult to visit the dentist due to a busy schedule or where medical infrastructure is limited.

This project was initiated when Professor Kim Hee-kyung of Ajou University Dental Hospital asked for collaboration in developing an AI model for caries diagnosis. It was necessary to solve problems such as data shortage, security issues, and data standardization that were encountered during the AI model development process.

SuperbAI completed the initial AI model within three weeks of the project’s start. The diagnostic criteria of the four participating dental specialists were reflected in the model, and the model performance was continuously improved based on the specialists’ feedback to increase accuracy. The insufficient caries data was supplemented through data augmentation technology based on generative AI.

We also carried out standardization work to minimize the quality variability of images taken with smartphone cameras. We safely processed medical data with the advanced security system within the platform, and we also established data collection guidelines and continuously collected insufficient data.

The newly developed model recorded high levels in 'precision', which is the rate at which actual cavities were accurately detected; 'recall', which is the rate at which cavities were found out among all cavities; and 'F-1 score', which is the harmonic mean of precision and recall. By establishing a platform that can provide diagnosis results in real time, the technology has been implemented at a level that can be used in actual fields.

Professor Hee-kyung Kim of Ajou University Dental Hospital said, “It is difficult to standardize medical diagnoses because they involve the subjective diagnosis of doctors,” and explained, “Because the project manager in charge learned dental terminology and actively reflected feedback, we were able to achieve a collaborative process without conflicts in expertise.”

Superb AI CEO Hyunsoo Kim said, “We can create the first protocol of the desired AI model in just three weeks and continuously improve it, so we can quickly confirm performance from the proof-of-concept (PoC) stage before proceeding with a large-scale project.” He added, “We will continue to contribute to the advancement of digital healthcare technology with our innovative technologies and platforms.”


  • See more related articles