TABLE 7.

Comparative table models used in computer vision.

Model	Key applications	Strengths	Weaknesses
CNN (Chua, 1997)	Component identification, defect detection	- High accuracy in image classification and object detection - Ability to learn complex patterns	- Requires large datasets - High computational cost
You Only Look Once (YOLO) (Jiang et al., 2022)	Real-time object detection	- Real-time processing - Single pass for object detection	- Lower accuracy for small objects - Less effective for highly cluttered scenes
Single Shot MultiBox Detector (SSD) (Liu et al., 2016)		- Faster than R-CNN - Balances speed and accuracy	- Lower accuracy compared to R-CNN for small objects
R-CNN (Zhang et al., 2018)	Object detection and segmentation	- High accuracy in object detection and segmentation - Effective for complex scenes	- Slower processing speed - High computational requirements
Fast R-CNN (Girshick, 2015)		- Faster than R-CNN- High accuracy in object detection and classification	- Still slower than YOLO and SSD - High computational cost
RetinaNet (Wang et al., 2019)		- High accuracy - Efficient in detecting objects of varying sizes	- Higher computational cost - Complex training process
Mask R-CNN (Vuola et al., 2019)		- Adds instance segmentation to object detection - High accuracy in detecting and segmenting objects	- Slower processing speed - High computational requirements