Interpretation of Human Actions from Live Video using Deep Learning Techniques

  • N. Prasath, R. Abijoy


The human monitoring system is one of the greatest challenges faced by various researchers in all these
years. There are various methods and techniques proposed by differerent professionals who work in this
area. The activity is monitored using the camera and the detection is done using neural networks. Every
method implies different results based on their characteristic features included in the model. There are
different approaches including the yolov3, object detection algorithms and various neural networks and
infer different successful results. To improve the additional features along with the existing model, the work
is carried out with the deep learning techniques and computer vision tools and the results obtained shows
a comparatively greater efficiency and can be achieved using simpler process. The execution of the process
is achieved by deep learning techniques by constructing the convolutional neural networks and computer
vision. This allows the model in detection and monitoring of multiple human activities from the video. The
paper explains the execution of the model and the results obtained by employing this method.