Moving Object Detection in Video Streaming Using Improved DNN Algorithm

An efficient approach for MOD (Moving Object Detection) has been implemented in research work. In this proposed work, collect the video samples such as *.avi format. After that, it developed the frame differencing to divide the video into frame format or extraction. It has converted the RGB color image into three different mechanisms like red, green, and blue channels. It has identified the noise level in the uploading video. It has developed the Gaussian filtration method to remove the noise in the uploaded image frames. It has developed the SURF algorithm to extract the features in the key-points format. In research work using improved classification algorithm to detect the moving objects. The experimental result analysis using the MATLAB simulation tool used and proposed work parameters have achieved the Accuracy value 99 per cent, Precision value 0.96, Recall Value 0.875, Specificity value 0.99, FPR value 0.0067, FNR value 0.124, F1-measure value 0.972, and compared with the existing parameters.