School of Mathematical Sciences, National Engineering Research Center of Visual Technology, State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University, Beijing, China
Auto-generated profile, based on publicly available information
Chenxi Yang
School of Mathematical Sciences, National Engineering Research Center of Visual Technology, State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University, Beijing, China
No-Reference Image Quality Assessment (NR-IQA) aims to predict image quality scores consistent with human perception without relying on pristine reference images, serving as a crucial component in various visual tasks. Ensuring the robustness of NR-IQA methods is vital for reliable comparisons of different image processing techniques and consistent user experiences in recommendations. The attack methods for NR-IQA provide a powerful instrument to test the robustness of NR-IQA. However, current attack methods of NR-IQA heavily rely on the gradient of the NR-IQA model, leading to limitations when the gradient information is unavailable. In this paper, we present a pioneering query-based black box attack against NR-IQA methods. We propose the concept of score boundary and leverage an adaptive iterative approach with multiple score boundaries. Meanwhile, the initial attack directions are also designed to leverage the characteristics of the Human Visual System (HVS). Experiments show our method outperforms all compared state-of-the-art attack methods and is far ahead of previous black-box methods. The effective NR-IQA model DBCNN suffers a Spearman’s rank-order correlation coefficient (SROCC) decline of 0.6381 attacked by our method, revealing the vulnerability of NR-IQA models to black-box attacks. The proposed attack method also provides a potent tool for further exploration into NR-IQA robustness.
No-Reference Image Quality Assessment (NR-IQA) plays a crucial role in various real-world applications by predicting image quality scores without the need for reference images. Despite the impressive performance of deep learning-based NR-IQA models, they remain vulnerable to adversarial attacks, which introduce imperceptible perturbations to input images, causing significant changes in predicted scores. In this study, we explore the use of simple JPEG compression techniques, as well as their combination with norm regularization training, to defend against these adversarial attacks. Our results demonstrate that image compression is an effective method to enhance model robustness, and it can further improve the robustness of NR-IQA models when combined with appropriate training strategies. Since excessive compression may reduce performance on clean images, it is essential to strike a balance. This work provides valuable insights into designing effective image compression methods for NR-IQA models.
This paper presents the attempt of modeling input-output nonlinear block structure systems with dynamic disturbances and measurement noise from sampled data. A typical Hammerstein-Wiener model is applied to fit and predict the output response of the system. The dynamic disturbance output will be viewed as a time-varying sequence to be estimated. By extending the parameter and information vector, a discrete-time extended recursive least squares algorithm is proposed to identify both the system parameters and dynamic disturbance. A matrix forgetting factor is utilized to improve the estimation accuracy and the tracking performance of the proposed algorithm. By using the auxiliary model technique, the consistent estimation of model parameters can be guaranteed. Moreover, several asymptotic convergence conditions are also derived based on the stochastic statistical theory under persistent excitation condition. Finally, a numerical example with Monte-Carlo simulation is used to illustrate the merit and effectiveness of the proposed algorithm.
Deep neural networks have demonstrated remarkable success in No-Reference Image Quality Assessment (NR-IQA). However, recent researches highlight the vulnerability of NR-IQA models to subtle adversarial perturbations, leading to inconsistencies between model predictions and subjective ratings. Current adversarial attacks, however, focus on perturbing predicted scores of individual images, neglecting the crucial aspect of correlation coefficients for predicted scores within an entire image set. Meanwhile, it is important to note that the correlation, like ranking correlation, plays a significant role in NR-IQA tasks. To thoroughly explore the robustness of NR-IQA models, we introduce a new framework of correlation-error-based attacks that perturb both the correlation within an image set and score changes on individual images. In this work, the proposed attack framework is implemented with ranking-related correlation metrics like Spearman's Rank-Order Correlation Coefficient (SROCC) and prediction error-related metrics like Mean Squared Error (MSE), which is called SROCC-MSE-Attack (SMA). The SMA works in two stages, firstly optimizing target attack scores for the entire image set and then generating adversarial examples guided by these scores. The SMA method not only significantly disrupts the SROCC to negative values but also maintains a considerable change in the scores of individual images. Meanwhile, it exhibits state-of-the-art performance across metrics in different categories. The proposed attack framework offers a new perspective for examining the robustness of NR-IQA models and holds potential for application with other correlation metrics and prediction error indices.