Search

ABSTRACTIn today's social networking era, information security breaches are at an all-time high, especially when it comes to texts and images, which should come as no surprise. Integrity is therefore crucial for information sharing over an unprotected connection. As a result, image/text encryption is a crucial preprocessing step in today's information transmission. Sensitive information may be sent via image encryption, which facilitates a variety of procedures. Here the proposed method is a combination of the Modified Caesar Cipher and the Card Deck Shuffle Algorithm for image encryption. The Card Deck Shuffle Algorithm rearranges the pixels from the result of the Modified Caesar approach, which encrypts the pixel value of each image. The method employs an n + 8 variable bit key (n is frequently more than 18 in real-world applications), which needs more than [Formula: see text] brute force attacks to be successful. The method described here may be used to secure images in a variety of multimedia applications.

ABSTRACTThis paper presents a novel approach for encrypting images using a Sudoku as its encryption key. This algorithm uses both symmetric and asymmetric key cryptography. It works with any type of data, Sudoku size, and keyspace. The image undergoes the process of modified thresholding, using a pseudo-random number generated from a Sudoku as the threshold. This image is then padded with zeros or the average pixel values to ensure the dimensions are multiples of the Sudoku's size and the image rows are shuffled randomly. For each iteration, the image rows are shuffled, followed by the columns, and finally, the image is rotated clockwise by 90 degrees. The resultant image is highly encrypted and resilient to brute-forcing methods. The algorithm requires roughly 25 milliseconds per iteration for a colored square image of dimensions [Formula: see text] and has an NPCR value of 99.60% and a UACI value of 35.65%. The gargantuan keyspace offered by the Sudoku keys ensures obedience of Kirchoff's principle and Shannon's maxim.

In the 21st-century, data is as valuable as gold. Many data-centric applications are generating a vast amount of data. Businesses can use this generated data to pinpoint the various sources of problems, if any. In addition, the data can help enterprises to identify connections between what is happening in different areas, departments, and systems. However, having more data is not enough; the data should also be of high quality. For example, taking action based on unfamiliar evidence, speculative ideas, or observations could lead to the wastage of resources. Whereas using high-quality data will help achieve correct results. Synthetic data is artificially generated data. Synthetic data is generated by an algorithm and used to represent real-world data, test datasets, perform mathematical model validation, and, most importantly, for training of machine learning models. Synthetic data can also be used to preserve data privacy. It is considered a safe way to transfer sensitive data because it creates a transaction database that does not contain any confidential information. This paper compares the tabular synthetic data generation techniques using various datasets, viz. balanced datasets, unbalanced datasets, datasets with numerical attributes only, datasets with categorical attributes only and mixed datasets. The utility of the generated synthetic data is measured using the Propensity score metric and Cluster-Log metric. The main finding of this paper is that the Classification And Regression Tree (CART) model provides the best results for all types of datasets. At the same time, Generative Adversarial Networks (GANs) give subpar or mediocre results at best. This contradicts the common belief that GANs are the go-to models for producing synthetic data.

Distributed Denial of Service (DDoS) attack is one of the most dangerous attacks that result in bringing down the server(s) and it is essential to classify the exact attack to implement robust security measures. In this work, we present an approach for detecting the prominent DDoS attacks that can be carried over Transport Layer protocols. Four different levels are taken into consideration which helps to classify one of the 11 different attacks. A bigger problem is divided into smaller ones and then conquered. This approach, called TaxoDaCML - Taxonomy-based Divide and Conquer approach using ML minimizes computational cost and at the same time maintains the required accuracy. Results prove that our approach achieves 99.9% accuracy for DDoS attack detection and more than 85% for DDoS attack classification. Comparison of TaxoDaCML is done with the previous works and is found to perform better for DDoS attacks classification.