Related Experiment Video
Updated: Aug 14, 2026
05:49
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Hybrid artificial intelligence framework for patient emotion recognition in dentistry
Aryadi Aryadi1, Harris Gadih Pratomo1, Kanoksak Wattanachote2
1Faculty of Dentistry, Trisakti University, Jakarta, Indonesia.
Frontiers in Oral Health
|August 13, 2026
Summary
This study created a new dataset for recognizing patient emotions in dental visits and tested an AI framework. The hybrid AI model showed promise for improving automated emotion detection in dentistry.
Area of Science:
- Artificial Intelligence
- Natural Language Processing
- Dentistry
Background:
- Emotional communication is crucial in dentist-patient interactions.
- Limited resources exist for automated emotion recognition in dental settings.
Purpose of the Study:
- Develop an emotional dental consultation corpus.
- Evaluate a hybrid AI framework for automated patient emotion recognition.
Main Methods:
- Transcribed and annotated 2,160 patient utterances from 12 consultations into six emotion categories.
- Trained Bidirectional Long Short-Term Memory (BiLSTM), BERT, and RoBERTa models.
- Developed a hybrid framework using large language models (LLMs) for confidence re-ranking.
Main Results:
- Achieved high annotation reliability (κ = 0.89).
- Neutral, Fear, and Disgust were the most prevalent emotions.
- Hybrid framework (BiLSTM + Qwen 2.5) achieved the highest macro-F1 score (0.92), though not statistically significant.
Conclusions:
- Presents a novel emotion-annotated dental corpus.
- Demonstrates feasibility of a hybrid AI framework for dental emotion recognition.
- Findings suggest preliminary utility of LLM-based re-ranking for ambiguous utterances.