Deteksi Lesi Pra-Kanker Serviks Pada Citra Kolposkopi Menggunakan Convolutional Neural Network dengan Arsitektur YOLOv7

Nurqolbiah, Fatihani
Nurmaini, Siti
Saputra, Tommy

Jurnal Sistem Komputer dan Informatika (JSON)
Vol. 5 No. 2 (2023): Desember 2023

Pre-cancerous cervical lesions detection is crucial in the diagnosis and analysis of medical images. Because visual observations are weak, computer-based detection is needed. This research proposes a pre-cancerous cervical lesion detection model using a Convolutional Neural Network with the YOLOv7 architecture, capable of accurately detecting these lesions. The data used was 913 colposcopy image data from 200 cases. The dataset is divided into training and testing data, resulting in a detection model for pre-cancerous cervical lesions. The model achieves an mAP of 91.9%, precision of 87.7%, recall of 96%, and an F1-score of 93%. The study demonstrates that the performance of YOLOv7 indicates the model's ability to accurately detect pre-cancerous lesions in the cervix.