Object Detection in Real Time Video
Shailendra S. Aote · Bioscience Biotechnology Research Communications · 2020
In this modern era, it is essential for various brands and businesses to have a technology that will be helpful for them with respect to the promotion of their business.Existing logo detection research focuses on closed set scenarios.The logo domain is too large for these strategies and requires an open set approach.In this paper, a logo detection and recognition system is proposed which detects the logos in the video.The issue of speed and size is addressed by proposing a compressed convolution neural network model VGG16 in Machine Learning.The proposed model improves on size, speed, and solves degradation issues.VGG16 trains the model and further classifies the images in order to detect the images.In comparison to VGG16 the proposed model is 88.4% smaller in size and 23.86% faster in training time.Thus, this model helps recognizing the images from the video with higher accuracy.