Comparative Estimation of Occlusal Contact Points Obtained with 2 Different Digital Systems with Artificial Intelligence

---

Şahin C. H., Toğay A., Tamam E.

The 48th Annual EPA Congress and The 27th Scientific TPID Congress, Nevşehir, Turkey, 11 - 13 September 2025, pp.59, (Summary Text)

• Publication Type: Conference Paper / Summary Text
• City: Nevşehir
• Country: Turkey
• Page Numbers: pp.59
• Gazi University Affiliated: Yes

Abstract

OBJECTIVES: In this study, a comparative analysis of 5

different deep learning models was performed to predict

T-SCAN occlusal analysis values from CEREC digital dental

restoration images using the data recorded at the maximum

intercuspal position.

MATERIALS-METHODS: Occlusal recordings were obtained

from healthy 20 females and 20 males aged 18-25 at the

maximum intercuspal position. Records were saved as.jpeg

format and transferred to Adobe Photoshop CS6 program.

Data prediction was performed with CNN (Convolutional

Neural Network), ResNet-50 (Residual Network-50), Vision

Transformer (ViT), Pix2Pix GAN (Pixel-to-Pixel Generative

Adversarial Network) and CNN+Attention hybrid model on the

dataset created using 80 images and corresponding T-SCAN

data, and evaluated with MAE (Mean Absolute Error), RMSE

(Root Mean Square Error) and clinical accuracy metrics.

RESULTS: The results indicate that models including attention

mechanism tend to perform better. This study investigates

the predictability of T-SCAN values from CEREC images with

artificial intelligence.

Keywords: T-Scan, Cerec, Occlusal analyzers, Artificial

Intelligence