AI based data processing algorithm for video surveillance (review) (research paper 1)

ABSTRACT

1. increase in CSS ( Camera surveillance System ) will have problem in storage & transmission bandwidth

2. This research paper focuses on background image & and moving object in scene & after that it will be transferred to server side for further processing.

3. This reduces storage side and increases the performance

INTRODUCTION

1. Storage capacity & transmission bandwidht is saved by separating Region Of Interest(ROI) or moving object & background .

2. Problem is that nowdays camera transmit whole entire frames are transmitted which create the above mentioned problems.

3. So valid backgrounds & ROIs are contructed at camera side and sent to server side for processing.

CONCLUSION

1. In this research paper Convolutional Neural Networks (CNN) are used for processing video surveillance data.

2. CNN is applied on both camera side and server side.

3. In camera side noticed objects are segmented by Neural Network (NN)

4. At server side D-CNN is used to evaluate the performance at camera side

5. Total time taken by this method is around 1 sec and by YOLO method is arround 5 sec.

src ->sciencedirect.com/science/article/abs/pii/S..