Table 1 Performance metrics and training configurations of different AI models used for dental caries detection, detailing the model names, number of epochs, batch sizes, GPUs utilized, evaluation metrics, and optimization algorithms.

Model Name	No. of epoch	Batch size	GPU	Metrics	Optimizer	Rank (Avg)
YOLOv5s	204	32	Nvidia RTX 4090 24GB	Precision = 0.768, Recall = 0.743, mAP @ 0.5 IoU = 0.78	Adam	2.67
YOLOv8s	164	64	Nvidia RTX 4090 24GB	Precision = 0.828, Recall = 0.807, mAP @ 0.5 IoU = 0.841	Adam	1.0
YOLO11s	225	64	Nvidia RTX 4090 24GB	Precision = 0.776, Recall = 0.735, mAP @ 0.5 IoU = 0.812	Adam	2.33
SSD-MobileNet-v2-FPNLite-320	5000 steps	16	T4 (15.0 GB) Colab	mAP @ 0.5 IoU = 0.68	RMSprop	4.9
FasterRCNN	3000 steps	8	T4 (15.0 GB) Colab	Recall = 0.461, mAP @ 0.5 IoU = 0.306	SGD	4.5

YOLOv5s: You Only Look Once versions 5, YOLOv8s: You Only Look Once version 8, YOLOv11: You Only Look Once version 11, Faster R-CNN: Faster Region-Convolutional Neural Network, Map: Mean Average Precision, GPU: Graphical Processing Unit, SGD: Stochastic Gradient Descent, Adam: Adaptive Moment Estimation, IoU: Intersection over Union.

Quick links

Search