Extracting image features with data integrity and confidentiality by verifiable outsourcing computation in cloud computing

Due to rapid enhancement of digital communication in cloud paradigm, easier transmission & storage of the multimedia information in several platforms becomes challenging. The security of image information is vital since the images are considered as a major component of communication in cloud environment. The secret information is shared in the form of secured image which needs to be retrieved and send to user without losing the integrity and confidentiality of data. For this purpose, the proposed model is designed which employs feature extraction categorization process and transmitting extracted information securely via cryptographic process. Initially the input images are retrieved and parameter initialization is carried by bilinear matrix. An optimal feature extraction is carried using Rotational invariant Local Binary Pattern (RI-LBP) along with Enriched Shark smell optimization process for extracting features of secret information. E-IBE (Enhanced-Identity based encryption) is employed for private key generation followed by cryptographic process via Ensemble Improved Homomorphic Pailler and Quantized ElGammal Elliptic curve Cryptography (ECC) scheme. The decrypted outcome attained is then digitally verified by employing SHA3 verification model. Thus, retrieved data is provided to the user after validation in a secured manner. The simulation results are then observed by analyzing the proposed scheme performance on CIFAR-10 dataset &MNIST dataset attained outcomes are compared with traditional schemes to validate the enhancement of proposed model over other models. the performance is carried for various metrics like extraction accuracy, recall, precision F1-score, precision-recall curve, RoC curve, execution time, runtime & storage space of entire system.