The network intrusion detection system (NIDS) plays a crucial role as a security measure in addressing the increasing number of network threats. The majority of current research relies on feature-ready datasets that heavily depend on feature engineering. Conversely, the increasing complexity of network traffic and the ongoing evolution of attack techniques lead to a diminishing distinction between benign and malicious network behaviors. In this paper, we propose a novel end-to-end intrusion detection framework based on a contrastive learning approach. We design a hierarchical Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU) model to facilitate the automated extraction of spatiotemporal features from raw traffic data. The integration of contrastive learning amplifies the distinction between benign and malicious network traffic in the representation space. The proposed method exhibits enhanced detection capabilities for unknown attacks in comparison to the approaches trained using the cross-entropy loss function. Experiments are carried out on the public datasets CIC-IDS2017 and CSE-CIC-IDS2018, demonstrating that our method can attain a detection accuracy of 99.9% for known attacks, thus achieving state-of-the-art performance. For unknown attacks, a weighted recall rate of 95% can be achieved.

Most machine learning algorithms only have a good recognition rate on balanced datasets. However, in the field of malicious traffic identification, benign traffic on the network is far greater than malicious traffic, and the network traffic dataset is imbalanced, which makes the algorithm have a low identification rate for small categories of malicious traffic samples. This paper presents a traffic sample synthesizing model named Conditional Tabular Traffic Generative Adversarial Network (CTTGAN), which uses a Conditional Tabular Generative Adversarial Network (CTGAN) algorithm to expand the small category traffic samples and balance the dataset in order to improve the malicious traffic identification rate. The CTTGAN model expands and recognizes feature data, which meets the requirements of a machine learning algorithm for training and prediction data. The contributions of this paper are as follows: first, the small category samples are expanded and the traffic dataset is balanced; second, the storage cost and computational complexity are reduced compared to models using image data; third, discrete variables and continuous variables in traffic feature data are processed at the same time, and the data distribution is described well. The experimental results show that the recognition rate of the expanded samples is more than 0.99 in MLP, KNN and SVM algorithms. In addition, the recognition rate of the proposed CTTGAN model is better than the oversampling and undersampling schemes.

A typical Chinese remainder theorem (CRT)-based secret sharing (SS) scheme has been proposed by Asmuth and Bloom for several decades, with lower computation complexity compared to Shamir’s original polynomial-based SS. But when applied to images, CRT-based image secret sharing (CRTISS) shows many problems, such as lossy recovery, auxiliary encryption and extra parameters requirement. We analyze the characteristics of images and ISS, and propose a (k,n)-threshold CRTISS based on the Asmuth and Bloom’s scheme by sharing the high 7 bits of a grayscale secret pixel and embedding the least significant bit (LSB) into the random integer. The pixel values of a grayscale image are divided into two parts, which make it possible to share all the secret pixels with no expansion. Our method has the advantages of (k,n) threshold, lossless recovery and no auxiliary encryption. The parameters requirement is the same as that in the Asmuth and Bloom’s original method. Analysis and experiments are provided to validate the effectiveness of the proposed method.

Secret image sharing (SIS) has received increased attention from the research community because of its usefulness in multiparty secure computing, access control, blockchain distributive storage and other security-oriented applications. Prevention of fake and dishonest participants is a key issue that has spurred interest in practical applications of SIS. Unfortunately, most previous SIS schemes failed to detect and locate fake or dishonest participants. In this article, an SIS for a ( k,n )-threshold without pixel expansion is presented, which can detect and locate both fake and dishonest participants. Using a screening operation, the proposed approach fuses the benefits of polynomial-based SIS, visual cryptographic scheme (VCS) , and hash functions to authenticate separate participants both with and without a dealer. In addition, the proposed approach achieves lossless rebuilding of the secret image. Analyses and experiments are conducted in this study to establish the effectiveness of the presented approach.

Compared with Shamir’s original secret image sharing (SIS), the Chinese-remainder-theorem-based SIS (CRTSIS) generally has the advantages of a lower computation complexity, lossless recovery and no auxiliary encryption. However, general CRTSIS is neither perfect nor ideal, resulting in a narrower range of share pixels than that of secret pixels. In this paper, we propose a practical and lossless CRTSIS based on Asmuth and Bloom’s threshold algorithm. To adapt the original scheme for grayscale images, our scheme shares the high seven bits of each pixel and utilizes the least significant bit (LSB) matching technique to embed the LSBs into the random integer that is generated in the sharing phase. The chosen moduli are all greater than 255 and the share pixels are in the range of [0, 255] by a screening operation. The generated share pixel values are evenly distributed in the range of [0, 255] and the selection of $(k,n)$ threshold is much more flexible, which significantly improves the practicality of CRTSIS. Since color images in RGB mode are made up of three channels, it is easy to extend the scheme to color images. Theoretical analysis and experiments are given to validate the effectiveness of the proposed scheme.