NOVEL METHOD FOR HIDING INFORMATION DEPENDING ON FAST CONVOLUTION NEURAL NETWORK RESULTS

Abstract

Hiding important information into suitable cover is crucial task in keeping security information from any edit, replace and change for the purpose of protecting special communication from unwanted effects from the third parity. Hiding information can be defined as a science or art for hiding important information in digital format such as text, audio, video and image (message) into a suitable digital file (cover) into deterministic sequence to prevent the detection of the hidden information from other unwanted parts. The proposed method consist of many steps that integrates for the purpose of producing compact result, the first step in the proposed system related to encryption process of the important text that prepare to be transformed to the suitable distention by extracting suitable key from a guide image in deterministic way and then ciphering text, the second step related to the constructing fast convolution neural network and used this network in detecting decided objects like cars, persons, animals,….. and then determined the coordinates of the points of the bounding box that cover the object after that in the third step a suitable algorithm proposed for hiding encrypted text into the image out of the bounding box in calculated position which is depend on the suitable metrics (edge strong) to decide the accepted position for hiding the bits of the message. The results of hiding text into suitable cover image have good values from the accepted range of the peak signal to noise ratio as well as any person cannot be recognized any difference between input and output images.