Object Detection Project from Video by Machine Learning Model Source Code

Project Discription

As video object detection plays an important role in surveillance system. Contributing this field of video object action detection is done in this paper. The key idea is to separate the foregoing and background pixels in the frame by using Gaussian mixture modal. Here object pixel pattern are taken as the training feature input in the neural network. Trained neural network by these Gaussian mixture makes an efficient video object action detection. Values obtained from different evaluation parameters shows that proposed work reduce execution time by 3.91 time. Results shows that multiple actions are detect from the same trained neural network.  This work considered only the dynamics of the video sequence without considering spatial information.As video object detection plays an important role in surveillance system. Contributing this field of video object action detection is done in this paper. The key idea is to separate the foregoing and background pixels in the frame by using Gaussian mixture modal. Here object pixel pattern are taken as the training feature input in the neural network. Trained neural network by these Gaussian mixture makes an efficient video object action detection. Values obtained from different evaluation parameters shows that proposed work reduce execution time by 3.91 time. Results shows that multiple actions are detect from the same trained neural network.  This work considered only the dynamics of the video sequence without considering spatial information.

Project Sample Image