Enhancing security of cloud data through encryption with AES and fernet algorithm through Convolutional-Neural-Networks (CNN)

Abstract

Cloud as a storage in the recent technological development had been focused by researchers, since it offers more insight towards meta-data based security and safety along with techniques in encryption and decryption of messages. “Data” being a crucial and complicated means-of-information in current technological era, it has been majorly accessed and utilized for varied purposes (example: image storage/access) by people globally through ‘cloud computing’ via social platforms, personal data-storage, professional data-accumulation, research based studies, etc. Thus to protect data in cloud, especially the images, the current study developed the algorithm by combining ‘AES’ and ‘Fernet’ where double-level encryption with CNN Auto-Encoders. Thus by developing the model, the study aims to provide more secured cloud computing model than existing models. The original images as input are processed, encrypted/decrypted, converted into bitmap images as outputs that are decrypted by users with ‘key’ when needed. The study was a success and found to be effective in image encryption field with high RMSE (0.040206), less MSE-Loss (0.001616) and MAE (0.0266323) scores than estimated scores.