Object Detection and Screen Presence Time Estimation Using Opencv and Yolo Alogrithm

This paper describes about the object detection and object’s screen presence time estimating model which will run on any computer easily. Currently there are only models that will only detect the motion of the object alongside estimating screen time. But here we tried to propose a model that will detect the object that is present in the camera view and also estimate the amount of time the object is present in the camera view. For object detection there are many algorithms such as CNN (Convolutional Neural Network), R-CNN, Fast R-CNN, Faster R-CNN, SSD (Single Shot Detector), YOLO (You Only Look Once) etc. In the current model we gave preference to the speed of detecting the object along with the accuracy. So, we preferred using the Yolo algorithm among all the existing algorithms that can be used for the object detection. But yolo is only designed for using in the GPU based computers. So, in order to implement yolo algorithm in our normal CPU based computers we used OpenCV library such that real time object detection is also possible in Non-GPU computers. Our model estimates the time of screen presence of each object using python libraries such as pandas, time etc. As we used yolo as our object detection algorithm our model detects objects with 80-99% confidence.