Gurpreet Kaur

Bio: Gurpreet Kaur is an academic researcher. The author has contributed to research in topics: Digital watermarking & Digital Watermarking Alliance. The author has an hindex of 1, co-authored 1 publications receiving 27 citations.

Image Watermarking Using LSB (Least Significant Bit)

Abstract: With the rapid development and wide use of Internet, information transmission faces a big challenge of security. People need a safe and secured way to transmit information. Digital watermarking is a technique of data hiding, which provide security of data. This paper presents a watermarking technique which least significant bits (LSB), its steps and its process with matlab images.

Invisible Backdoor Attacks on Deep Neural Networks via Steganography and Regularization

Abstract: Deep neural networks (DNNs) have been proven vulnerable to backdoor attacks, where hidden features (patterns) trained to a normal model, which is only activated by some specific input (called triggers), trick the model into producing unexpected behavior. In this paper, we create covert and scattered triggers for backdoor attacks, invisible backdoors, where triggers can fool both DNN models and human inspection. We apply our invisible backdoors through two state-of-the-art methods of embedding triggers for backdoor attacks. The first approach on Badnets embeds the trigger into DNNs through steganography. The second approach of a trojan attack uses two types of additional regularization terms to generate the triggers with irregular shape and size. We use the Attack Success Rate and Functionality to measure the performance of our attacks. We introduce two novel definitions of invisibility for human perception; one is conceptualized by the Perceptual Adversarial Similarity Score (PASS) and the other is Learned Perceptual Image Patch Similarity (LPIPS). We show that the proposed invisible backdoors can be fairly effective across various DNN models as well as four datasets MNIST, CIFAR-10, CIFAR-100, and GTSRB, by measuring their attack success rates for the adversary, functionality for the normal users, and invisibility scores for the administrators. We finally argue that the proposed invisible backdoor attacks can effectively thwart the state-of-the-art trojan backdoor detection approaches, such as Neural Cleanse and TABOR.

Invisible Backdoor Attacks on Deep Neural Networks Via Steganography and Regularization

Abstract: Deep neural networks (DNNs) have been proven vulnerable to backdoor attacks, where hidden features (patterns) trained to a normal model, which is only activated by some specific input (called triggers), trick the model into producing unexpected behavior. In this article, we create covert and scattered triggers for backdoor attacks, invisible backdoors , where triggers can fool both DNN models and human inspection. We apply our invisible backdoors through two state-of-the-art methods of embedding triggers for backdoor attacks. The first approach on Badnets embeds the trigger into DNNs through steganography. The second approach of a trojan attack uses two types of additional regularization terms to generate the triggers with irregular shape and size. We use the Attack Success Rate and Functionality to measure the performance of our attacks. We introduce two novel definitions of invisibility for human perception; one is conceptualized by the Perceptual Adversarial Similarity Score (PASS) and the other is Learned Perceptual Image Patch Similarity (LPIPS). We show that the proposed invisible backdoors can be fairly effective across various DNN models as well as four datasets MNIST, CIFAR-10, CIFAR-100, and GTSRB, by measuring their attack success rates for the adversary, functionality for the normal users, and invisibility scores for the administrators. We finally argue that the proposed invisible backdoor attacks can effectively thwart the state-of-the-art trojan backdoor detection approaches.

Comparative analysis of LSB, DCT and DWT for Digital Watermarking

Abstract: In one's day to day life; internet is growing and became an important part. Digital content can easily be downloaded, copied or edited. Digital content can be secured by various ways. Digital Watermarking is one of the methods for the protection of Digital Content. Digital Watermarking is a method by which data can be secured by hiding data in any image which can work as a carrier image. The carrier image is also known as cover image. Watermarking is an interactive method to protect and identify the digital data. It permits various types of watermarks to be hidden in digital data e.g. image, audio and video. In this work, watermarking has been done using LSB technique, DCT transform and DWT transform. These techniques are compared on the basis of peak signal to noise ratio (PSNR) and normalized correlation (NC). Parameters are compared for various noises like Gaussian noise, Poisson noise, Salt and Pepper noise and Speckle noise.

Digital video watermarking using modified lsb and dct technique

Abstract: From the previous millennium there has been a significant change from the analog to digital world. In today's ti me Audio cd's, Internet and DVD's are more widespread than before. Though the film and music content owners are relu ctant to release digital contend. This situation is because if this digital content is left unprotected, there are high chances of it being copied rapidly, on a large scale, with no limitations on the number of copies and it can be easily distributed via Internet. [3]As there is a rapid development in the current Information Technology, electronic publishing like the distribution of digitized images and videos, are becoming more and more popular. So there is a need of protecting as well as authenticating 3 dimensional data. Copyright protection is an important issue for electronic publishing. This paper presents the different video watermarking techniques. It provides a review on various available algorithms. In addition to it, focus on Combined Least Significant Bit (LSB) and Discrete Cosine Transform(DCT) Technique and also LSB with modifications Technique.

A Review on Digital Image Watermarking and Its Techniques

Abstract: The enormous popularity of internet offers various multimedia resources through various digital networks. These multimedia resources or digital media should be protected against various unauthorized attacks so as to use them for profit or security. Digital Watermarking is a way of protecting the digital media from unauthorized usage.This paper is a review on the Watermarking process, Types of watermarks, Various Watermarking Techniques and Applications of Watermarking. 