Integration of object detection and face recognition in real time
Main Article Content
Abstract
This paper presents a new system for integrating real-time object detection and face recognition technique.
Firstly, object detection technology was used by model YOLOv3, which is trained on the darknet53 network that uses COCO dataset.
The YOLOv3 model can detect 80 different classes of objects.
Based on that, we will be able to detect people, buses, cats, dogs, cars, and more.
while Haar cascade classifier and Local Binary Patterns Histogram (LBPH) algorithm were used to recognize faces and identify the names of people and objects on the camera in real time.
Experiments showed the effectiveness of the proposed system for real-time object detection using YOLOv3-416 achieving a mAP-50 of 55.3, with an accuracy in the range of 81-92%.
While the accuracy of real time face recognition using Haar cascade classifier and LBPH algorithm was 83%.