AI model from Ateneo Lab helps dentists spot infections with precision

A global research team led by the Ateneo Laboratory for Intelligent Visual Environments (ALIVE) is setting a new standard in dental diagnostics, thanks to a breakthrough in artificial intelligence. The group's deep learning model, developed in collaboration with Taiwan-based researchers, can now identify teeth and sinus structures in dental X-rays with 98.2% accuracy—a game-changer for a condition many doctors still miss.

At the heart of the innovation is YOLO 11n, a refined version of the well-known "You Only Look Once" object detection algorithm. Trained on thousands of dental panoramic radiograph (DPR) images, the model can detect subtle anatomical clues that point to odontogenic sinusitis—a condition frequently misdiagnosed as general sinusitis.

"It's easy to miss," the ALIVE team said in a release. "The symptoms mimic general sinusitis: nasal congestion, foul nasal discharge, and sometimes toothache. But only about a third of patients actually feel that tooth pain."

Don't miss: Dignity, dentures, and data: How Haleon went hyperlocal to restore smiles

Often, proper diagnosis requires input from both dentists and ENT specialists, a hurdle that delays treatment. Left unchecked, odontogenic sinusitis can lead to infections spreading to the eyes, face, or even brain.

The YOLO 11n model stands out not only for its accuracy but also for its efficiency. By identifying affected areas in real time and in a single pass, the tool reduces reliance on CT scans—currently the diagnostic gold standard—and limits patients' exposure to radiation. This is especially impactful in under-resourced healthcare settings where advanced imaging may not be readily available.

Patricia Angela R. Abu, head of ALIVE, noted that the system is more than just a clever tool—it's a potential equaliser. "In areas where access to specialist care is limited, this technology can support earlier intervention and improved patient outcomes," she said.

This breakthrough is part of a broader trend of AI integration into clinical environments. Beyond dentistry, deep learning continues to support physicians across specialisms, helping interpret images faster and more reliably than ever before.

The findings, which Abu and colleagues from Chang Gung Memorial Hospital, National Cheng Kung University, Chung Yuan Christian University, and Ming Chi University of Technology have published in Bioengineering, mark a critical step forward in dental AI.

With further validation, this assistant could soon become a fixture in dental clinics—quietly reviewing X-rays, catching what the eye might miss, and ensuring that no condition goes untreated.

Related articles:
J.P. Morgan bets on the Philippines in regional AI payments push
Report: B2B marketers still struggle to reach the right audiences, but AI may help
Pepsodent raises dental health awareness with new tele-dental consultations