Penerapan Deep Learning dalam Analisis Citra Gigi

Supiyandi Supiyandi (Universitas Pembangunan Panca Budi)
Wahyu Eka Judistira (Universitas Islam Negeri Sumatera Utara)
Sepriana Nurliani (Universitas Islam Negeri Sumatera Utara)
Rondi Sahputra Darmono (Universitas Islam Negeri Sumatera Utara)
Inneke Putri (Universitas Islam Negeri Sumatera Utara)

Jurnal Pendidikan dan Ilmu Sosial
Vol. 2 No. 4 (2024): Oktober : JURNAL PENDIDIKAN DAN ILMU SOSIAL

Testing in dental medical recognition and recording is still done manually, causing it to take a long time. In this study, an object detection method was applied to assist doctors in identifying patient conditions. Convolutional Neural Network (CNN) method was trained with an intraoral image dataset that includes five categories of tooth conditions: normal, filling, caries, and residual roots. CNN performance evaluation was conducted using evaluation metrics, and the results showed that the best CNN model achieved an mAP of 84% and a testing accuracy of 82%. This research successfully achieved its main goal, which is to build a reliable deep learning model for dental disease detection and recognition in humans.