Evaluation of Synthetic Panoramic Images by Dental Professionals, Students and AI models
2nd International Symposium on AI-Driven Engineering Systems, ISADES 2026, Hybrid, Mbale, Uganda, 19 - 20 Haziran 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/isades69945.2026.11608179
- Basıldığı Şehir: Hybrid, Mbale
- Basıldığı Ülke: Uganda
- Anahtar Kelimeler: Generative adversarial networks, Image synthesis, large language models, Panoramic dental radiographs
- Çanakkale Onsekiz Mart Üniversitesi Adresli: Evet
Özet
Large language models are frequently used for the evaluation of health-related data due to their online accessibility and their ability to provide responses on a wide range of topics. Another rapidly advancing area of artificial intelligence, image generation, has recently become capable of producing highly realistic visual content. This study evaluates how dentistry academics, students, and LLM-based chatbots assess synthetic images produced by conditional generative adversarial models. Within the study, participants were asked to complete a survey consisting of four real images and randomly selected pairs of synthetic images generated as single-class and multi-class outputs using the pix2pixHD, SPADE, CLADE, SPADE2XGRNX, SPADE2XGRNX symmetry, and SPADE2XGRNX symmetryInBlocks models. The results indicated that dentistry academics correctly identified the real images; however, they were also misled by some synthetic images. Dental students reported that they were not yet confident in distinguishing between real and synthetic images. As expected, chatbot systems utilizing large language models produced responses that were essentially random regarding the realism of the images.