Performance of Deep Convolutional Neural Networks for Motion Detection in Video Frames

In this paper, the performance of deep Convolutional Neural Networks (CNNs) with the number of different layers has been applied to classify video frames. The applied approach emphasizes on the health of workers and shows deep CNN architectures accurately learn features of objects as opposed to more shallow CNN architecture. Finally, the results indicate that deeper convolutional neural network is more efficient and this method is useful when there are a lot of data available.

کلیدواژه ها:

Convolutional Neural Network ، Motion Detection ، Hidden Layer ، Fully Connected Layer ، Rectified Linear Unit Layer ، Convolutional Layer.