An Efficient Object Detection Using YOLOV8 and SMCNN in Real Time Video Surveillance

In computer Vision, the ability to determine the presence of a certain object and locate it in places such as images and videos is referred to as object detection. It is In-used to detect a lot of objects in a session and draw boxes around them. The methods such as YOLO, SSD, and Faster R-CNN are common to applications such as self-driving cars and intelligent object monitoring systems. This paper proposes a model method using YOLO V8 along with spider monkey and CNN(SM-CNN). The agile and dynamic movement patterns of spider monkeys act as motivation for object detection models to track and identify fast moving objects in real time. Their unique modes of locomotion enhance certain datasets that train artificial intelligence, making it more efficient in locating animals and objects of similar nature within different environments. Convolutional Neural Networks (CNNs), advanced frameworks in deep learning, are primarily used for processing geospatial data, especially images, which often involve the extraction of critical features through several convolution layers. In object detection, CNNs enable accurate recognition and localization of objects by learning hierarchical patterns, making them essential for tasks like image classification and segmentation. Through numerous experiments, it has been shown that the suggested SM-CNN model is effective in object detection, achieving an average accuracy of 96.41% and reporting better accuracy than the older methods in F-measure, precision and error rate, thus proving its worth in practical situations.