arXiv subjects

Alexander Gushchin

Publications and source records attributed to Alexander Gushchin.

6 recordsLinked to original sources

arXiv2026

NIC-RobustBench: A Comprehensive Open-Source Toolkit for Neural Image Compression and Robustness Analysis

Neural image compression (NIC) is increasingly used in computer vision pipelines, as learning-based models are able to surpass traditional algorithms in compression efficiency. However, learned codecs can be unstable and vulnerable to adversarial attacks: small perturbations may cause severe reconstruction artifacts or indirectly break downstream models. Despite these risks, most NIC benchmarks only emphasize rate-distortion (RD) performance, focusing on model efficiency in safe, non-adversarial scenarios, while NIC robustness studies cover only specific codecs and attacks. To fill this gap, we introduce \textbf{NIC-RobustBench}, an open-source benchmark and evaluation framework for adversarial robustness of NIC methods. The benchmark integrates 8 attacks, 9 defense strategies, standard RD metrics, a large and extensible set of codecs, and tools for assessing both the robustness of the compression model and impact on downstream tasks. Using NIC-RobustBench, we provide a broad empirical study of modern NICs and defenses in adversarial scenarios, highlighting failure modes, least and most resilient architectures, and other insights into NIC robustness. Our code is available online at https://github.com/msu-video-group/NIC-RobustBench.

eess.IV↗

arXiv2025

Guardians of Image Quality: Benchmarking Defenses Against Adversarial Attacks on Image Quality Metrics

In the field of Image Quality Assessment (IQA), the adversarial robustness of the metrics poses a critical concern. This paper presents a comprehensive benchmarking study of various defense mechanisms in response to the rise in adversarial attacks on IQA. We systematically evaluate 25 defense strategies, including adversarial purification, adversarial training, and certified robustness methods. We applied 14 adversarial attack algorithms of various types in both non-adaptive and adaptive settings and tested these defenses against them. We analyze the differences between defenses and their applicability to IQA tasks, considering that they should preserve IQA scores and image quality. The proposed benchmark aims to guide future developments and accepts submissions of new methods, with the latest results available online: https://videoprocessing.ai/benchmarks/iqa-defenses.html.

cs.CV↗

arXiv2023

Video compression dataset and benchmark of learning-based video-quality metrics

Video-quality measurement is a critical task in video processing. Nowadays, many implementations of new encoding standards - such as AV1, VVC, and LCEVC - use deep-learning-based decoding algorithms with perceptual metrics that serve as optimization objectives. But investigations of the performance of modern video- and image-quality metrics commonly employ videos compressed using older standards, such as AVC. In this paper, we present a new benchmark for video-quality metrics that evaluates video compression. It is based on a new dataset consisting of about 2,500 streams encoded using different standards, including AVC, HEVC, AV1, VP9, and VVC. Subjective scores were collected using crowdsourced pairwise comparisons. The list of evaluated metrics includes recent ones based on machine learning and neural networks. The results demonstrate that new no-reference metrics exhibit a high correlation with subjective quality and approach the capability of top full-reference metrics.

cs.CV↗

arXiv2021

Shot boundary detection method based on a new extensive dataset and mixed features

Shot boundary detection in video is one of the key stages of video data processing. A new method for shot boundary detection based on several video features, such as color histograms and object boundaries, has been proposed. The developed algorithm was tested on the open BBC Planet Earth [1] and RAI [2] datasets, and the MSU CC datasets, based on videos used in the video codec comparison conducted at MSU, as well as videos from the IBM set, were also plotted. The total dataset for algorithm development and testing exceeded the known TRECVID datasets. Based on the test results, the proposed algorithm for scene change detection outperformed its counterparts with a final F-score of 0.9794.

cs.CV↗

arXiv2019

Drift Estimation for a Lévy-Driven Ornstein-Uhlenbeck Process with Heavy Tails

We consider the problem of estimation of the drift parameter of an ergodic Ornstein--Uhlenbeck type process driven by a Lévy process with heavy tails. The process is observed continuously on a long time interval $[0,T]$, $T\to\infty$. We prove that the statistical model is locally asymptotic mixed normal and the maximum likelihood estimator is asymptotically efficient.

math.ST↗

arXiv2014

Processes that can be embedded in a geometric Brownian motion

The main result is a counterpart of the theorem of Monroe [\emph{Ann. Probability} \textbf{6} (1978) 42--56] for a geometric Brownian motion: A process is equivalent to a time change of a geometric Brownian motion if and only if it is a nonnegative supermartingale. We also provide a link between our main result and Monroe [\emph{Ann. Math. Statist.} \textbf{43} (1972) 1293--1311]. This is based on the concept of a \emph{minimal} stopping time, which is characterised in Monroe [\emph{Ann. Math. Statist.} \textbf{43} (1972) 1293--1311] and Cox and Hobson [\emph{Probab. Theory Related Fields} \textbf{135} (2006) 395--414] in the Brownian case. We finally suggest a sufficient condition for minimality (for the processes other than a Brownian motion) complementing the discussion in the aforementioned papers.

math.PR↗