Perbandingan Deteksi Letak Polip pada Citra Colonoscopy menggunakan CNN dengan Arsitektur RetinaNet

ABSTRAKPenyakit kanker kolorektal diawali munculnya polip pada usus besar yang dapat berubah menjadi tumor ganas dan menimbulkan kanker. Sehingga diperlukan screening terhadap menggunakan colonoscopy. Menurut penelitian sekitar 26% terlewat saat prosedur Pada ini dilakukan implementasi Convolutional Neural Network (CNN) dengan arsitektur RetinaNet untuk mendeteksi letak citra Perbandingan 3 jenis yaitu ResNet-50, ResNet-101, ResNet-152 sebagai backbone RetinaNet. Model terbaik berdasarkan metrik Intersection over Union (IoU) adalah model (Backbone = ResNet-50) tanpa data augmentation nilai 0.8415. Sedangkan Average Precision (AP) ResNet-101) AP25 0.9308, AP50 =0.9039, AP75 0.6985.Kata kunci: polip, colonoscopy, (CNN), ABSTRACTColorectal cancer always begins with the appearance of polyps in colon which can turn into malignant tumors and cause cancer. Therefore, it is necessary to screen large intestine using However, according studies, about are missed during colonoscopy procedures. In this study, a architecture was implemented detect location images. Comparisons were made on types architecture, namely ResNet-152. From evaluation results, best based metric without value While values 0.9039, 0.6985.Keywords: polyp,