Prediction of Human Motion Detection in Video Surveillance Environment Using Tensor Flow
In recent year, Deep learning concepts are used to investigate image processing and machine learning applications. As the same experience we use deep learning prediction method for finding moving object and human motion in video surveillance environment. Deep learning approach is used to detect, classify and find the moving object and classify the person in captured videos. For finding human, a set of image processing steps are applied in moving areas that contain moving objects. The selected area are applied convolutional neural network classifier which has various layers and available in Google Tensor Flow learning tool. In this approach, shaking videos, low resolution videos are taken into account and video dataset examined. The experimental studies says that the relationship between moving object and video surveillance dataset has been analyzed and prediction performance has noted as 85% accuracy and from 65 videos.
Keywords: Deep Learning, Human Motion Detection, Convolutional Neural Network Classifier, Google Tensor Flow, Image Processing.