Second mesio-buccal canal segmentation with YOLOv5 using CBCT images

The objective of this study is to use a deep learning model based on CNN architecture to detect the second mesiobuccal (MB2) canals, which are seen as a variation in maxillary molars root canals. The CBCT images from the radiology archive of the Oral and Maxillofacial Radiology Department at Ankara University were used. You Only Look Once v5(YOLOv5) is a single-stage target recognition algorithm consisting of three main components: Backbone, Neck, and Head. For the segmentation task, the YOLOv5x architecture was chosen due to its high accuracy, speed and lightness, and DL; Using the Python open-source programming language. After the model was trained on axial slices of CBCT images with 500 epochs and a learning rate of 0.01, the fine-tuned deep learning model was tested on an independent test dataset, and the best model was recorded To evaluate the performance of the YOLOv5x model, precision, recall, F1 score, and mAP values were calculated at 0.5 Inter-over union(IoU) value using the confusion matrix's True Positive(TP), False Positive(FP) and False Negative(FN) parameters. MB2 canal was determined from axial slices of CBCT images of maxillary molar teeth using a CNN YOLOv5x-based AI software in this work. In this study, MB2 channel diagnosis was detected using different FOVs such as 8x8 cm2,12x8 cm2 and 15x12 cm2. Based on the developed DL models, the AI model gave successful results for MB2 segmentation in CBCT images. The MB2 canal of maxillary molars can be detected and segmented in CBCT images using deep learning-based artificial intelligence systems.