Year: 2024
Type: conference-paper
Abstract: Surveillance video is an integral part of our modern civilization, and such video data helps us continuously monitor different activities occurring around our society's private and public places.The captured surveillance video data needs huge storage space since the video data are bulky, and such data is further stored with some cryptographic approach to ensure security.So, in this paper, the authors have addressed how to effectively store such a massive volume of data in a cloud environment in a real-time scenario.Firstly, we have identified the foreground content, which is visually more sensitive, through the YOLO model; subsequently, we have segregated the video data into foreground-and backgroundcontent.Thereafter, the standard MPEG video compression was executed independently on the obtained segregated video components.To realize the encryption process in real-time, we have followed full encryption on the foreground part of the compressed video, whereas partial encryption is deployed on the background part.In the enciphering process, we have adopted AES-GCM and logistic map-based generated values to perform the encryption process.We have implemented both the compression and encryption process on the Raspberry Pi platform, and subsequently, the enciphered data is stored on the AWS cloud environment.The suggested framework is found realistic for storing surveillance video data securely in the cloud environment.
Authors: Shivam Gupta, Bhaskar Biswas, Arup Kumar Pal
Institution: Indian Institute of Technology Dhanbad
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4411099040