Deep learning-based identification, classification and segmentation of infraosseous periodontal defects using a YOLOv8 neural network
Xukai Gu, Petra Orosz, Kristof Somodi, Xinda Li, Lei Yang, Hong Yang, Christoph Ramseier, Daniel Palkovics · 2026
Read Deep learning-based identification, classification and segmentation of infraosseous periodontal defects using a YOLOv8 neural network by Xukai Gu, Petra Orosz, Kristof Somodi, Xinda Li, Lei Yang, Hong Yang, Christoph Ramseier, Daniel Palkovics (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13576134.txt.
Description
Xukai Gu leads a 2026 study training a neural network to find and outline bone defects around teeth on dental X-rays. These infraosseous defects, pockets where gum disease has eaten bone beside a root, guide prognosis and surgery planning. Overlapping structures on flat X-rays make them easy to misread.
The team trained a YOLOv8 network on 329 X-rays from Semmelweis University, each marked by hand. The model scored 0.8078 on the ranking accuracy measure, with a mid-range classification score. Its outlines overlapped expert markings well on average, strongest for one-wall, three-wall, and four-wall defects. The authors judge the model capable yet short of clinical polish. A larger, cleaner training set is their stated next step.
Catalog record
| CREATOR | Xukai Gu, Petra Orosz, Kristof Somodi, Xinda Li, Lei Yang, Hong Yang, Christoph Ramseier, Daniel Palkovics |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38498 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |