الفهرس 
Chapter (1)Only 14 pages are availabe for public view
 |  | from | 155 |  |  |
| | | from | 155 | | |
Abstract
Object recognition is a component of artificial intelligence which is a fast rising field of computer science. Object recognition is a procedure used to classify objects in videos or images, and real-time applications. Human detect variation of things in images directly without taking any struggle, even if the variance in image sizes or even images are translated or rotated. The objective of object recognition is to replicate this human intelligence using a computer algorithms. Convolutional neural network (CNN) is one of the key components of deep learning which has targeted a significant evolution in the state art of deep learning systems, specifically in the Human Activity Recognition (HAR) field. In this study, a novel CNN deep learning model has been proposed in order to increase the recognition accuracy of different human activities. The proposed model uses raw data acquired from a set of inertial sensor and exploring numerous combinations of human activities; sitting, standing, jogging, both up and downstairs, and walking. Experimental results show that the proposed CNN has achieved about 97.5% with respect to the NAdam optimizer which would be considered as the most effectively recognizer compared to other deep learning architectures. Moreover, this study introduces the implementation of modern YOLO algorithms (YOLOv3, YOLOv4, and YOLOv5) for multiclass 3D object detection and recognition. All YOLO algorithms have been tested according to a very large scaled dataset (Pascal voc dataset). Experimental results demonstrate that the YOLOv3 has targeted mAP of 77%, and the total running time was almost 8 hours. Moreover, in YOLOv4, it has targeted mAP of 55% and the total running time nearly 7 hours. In addition, YOLOv5 has established the mAP of 48% and the total running time was about 3 hours. Finally, a modified version of YOLOv5 has been proposed in the state of the art of optimizing its hyper-parameters and layering system. Accordingly, the Map scored about 55% with 3 hours running time. The final conclusions of this study have demonstrated that the proposed modified YOLOv5 has scored the least processing time.