ARCHIVES

Abstract: Us humans can detect and recognise objects very quickly by using the senses of sight and touch. But when it comes to computers or machines, it is a very challenging one. Object detection is a computer vision technique for locating objects in images or videos. In other words, object detection is a supervised learning problem, which means that train our models on labelled examples. Each frame in the training dataset must be accompanied with a file that includes the boundaries and classes of the objects. It is useful in many fields like automatic car, military, etc. There are numerous calculations for identifying the items. The popular exciting algorithm of object detection are CNN, RCNN. But the main disadvantages are the accuracy rate is very low, when it comes to real time applications.To overcome the traditional CNN algorithm, additional features were added to the algorithm and it was named as YOLO. YOLO algorithm employs Convolutional Neural Network(CNN) to detect objects in real time. As the name suggests, the algorithm requires only a single forward propagation through a neural network to detect objects. This means the prediction in the entire frame is done in a single algorithm run. YOLO algorithm accuracy rate and time complexity areefficient.In YOLO algorithm, YOLOv5 is a recent version. While comparing to other version, YOLOv5 is an efficient in application like object identification, fault detection, etc. This paper uses the YOLOv5 for recognizing images from video. It first split the videos into frames, that frames can be considered as images. Then it split the frames (i.e., images) by using the CNN (Convolutional Neural Network). By using the training dataset, the machine will analyse the objects in the frames. Then the output is displayed with the prediction of objects with highaccuracy.