An Approach to Cryptography Based on Neural Network Using Image for Public Key Generation

Keys play an important role in cryptography. A random number generator is often used to generate the key in cryptography. But there is no computation that really generates a perfectly random number. The use of images as key generating objects is one solution that can be used. In this digital era, the availability of images is very large and easy to obtain. Images have a value that text data does not have, which is rich in information. Each image has its own characteristics so the characteristics of the image are not general but very dependent on the model and object of the image used. In addition, using images as key generator objects will be very effective where keys do not need to be remembered and are safer in storage. This study aims to generate asymmetric keys using images and Single Layer Perceptron. The image will act as the public key. The method used to extract values from images is gray level co-occurrence matrix (GLCM). As for the private key generation process, the encryption process and the decryption process use a Single Layer Perceptron. GLCM feature extraction in the image can be used as an asymmetry key with the ideal number is ten nodes and achieving 100% accuracy