Haoxue Zhang

Haoxue Zhang

Auto-generated profile, based on publicly available information

Haoxue Zhang

School of Electronic and Information Engineering, Taiyuan University of Science and Technology

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 April 2025 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
LLLinjuan Li
Linjuan Li
GXGang Xie
Gang Xie
HZHaoxue Zhang
Haoxue Zhang
XXXinlin Xie
Xinlin Xie
Heng LiHeng Li
Heng Li
Remote sensing scene classification enables data-driven decisions for various applications, such as environmental monitoring, urban planning, and disaster management. However, deep learning models used for scene classification are highly vulnerable to adversarial samples, resulting in incorrect predictions and posing significant risks. While most current methods focus on improving adversarial robustness, they face a trade-off that compromises accuracy on clean, unperturbed images. To address this challenge, we utilized information theory by incorporating a mutual information (MI) representation (MIR) module, which allows the model to capture high-quality, robust features. Furthermore, a domain adversarial training (AdvT) strategy is applied to promote the learning of domain-invariant features, reducing the effect of distribution differences between clean images and adversarial samples. We propose a novel algorithm that accurately differentiates between clean and adversarial scenes by introducing the MI and domain adaptation-guided network (MIDANet). Extensive experiments demonstrate the effectiveness of our approach against adversarial attacks, revealing a positive correlation between adversarial perturbations and image information entropy, and a negative correlation with robust accuracy (RAcc).

19 September 2024 in Electronics
Verified
Journal Article
Open access
PDF file
LLLinjuan Li
Linjuan Li
HZHaoxue Zhang
Haoxue Zhang
GXGang Xie
Gang Xie
ZZZhaoxiang Zhang
Zhaoxiang Zhang
Deep learning models excel in interpreting the exponentially growing amounts of remote sensing data; however, they are susceptible to deception and spoofing by adversarial samples, posing catastrophic threats. The existing methods to combat adversarial samples have limited performance in robustness and efficiency, particularly in complex remote sensing scenarios. To tackle these challenges, an unsupervised domain adaptation algorithm is proposed for the accurate identification of clean images and adversarial samples by exploring a robust generative adversarial classification network that can harmonize the features between clean images and adversarial samples to minimize distribution discrepancies. Furthermore, linear polynomial loss as a replacement for cross-entropy loss is integrated to guide robust representation learning. Additionally, we leverage the fast gradient sign method (FGSM) and projected gradient descent (PGD) algorithms to generate adversarial samples with varying perturbation amplitudes to assess model robustness. A series of experiments was performed on the RSSCN7 dataset and SIRI-WHU dataset. Our experimental results illustrate that the proposed algorithm performs exceptionally well in classifying clean images while demonstrating robustness against adversarial perturbations.

7 July 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
JTJuan Tian
Juan Tian
YXYihan Xue
Yihan Xue
HZHaoxue Zhang
Haoxue Zhang
JLJinpeng Liu
Jinpeng Liu
GXGang Xie
Gang Xie
In industrial field, the data collected for fault diagnosis are often affected by the simultaneous presence of noise interference, varying working conditions, and limited fault samples, which lead to severe degradation in diagnostic performance. To address these challenges, a time–frequency guided dual stage domain adaptation method is proposed based on semi-supervised fault diagnosis paradigm. At first, a dual-branch feature extraction block is introduced to explore time domain and frequency domain representations separately. A bottleneck fusion mechanism is then developed to obtain compact and discriminative joint features. Guided by time–frequency prediction consistency, a dynamic pseudo label generation strategy is designed to effectively utilize unlabeled data, while enabling reliable extraction of latent information under complex working conditions. Furthermore, a dual-stage domain adaptation strategy is established to reduce domain discrepancy. In the first stage, subspace adversarial learning is applied to align time domain and frequency domain feature distributions, achieving local distribution alignment. In the second stage, global distribution alignment is achieved in the fused feature space by incorporating multi-kernel maximum mean discrepancy. Experimental results on bearing and planetary gearbox datasets show that the proposed method significantly improves cross-domain generalization performance compared to the existing approaches; moreover, the robust representation of the limited data under complex working conditions is also enhanced.

18 June 2025 in International Journal of Computers and Applications
Verified
Journal Article
Closed access
LLLinjuan Li
Linjuan Li
GXGang Xie
Gang Xie
HZHaoxue Zhang
Haoxue Zhang
JTJuan Tian
Juan Tian