arXiv subjects

Ye Zheng

Publications and source records attributed to Ye Zheng.

At least 19 recordsLinked to original sources

arXiv2026

Two-Stage Teacher-Student Reliable Prior Learning for Robust Underwater Image Enhancement

Underwater image enhancement (UIE) aims to recover clear images from observations affected by wavelength-dependent absorption, scattering, and spatially nonuniform degradation. Although existing generative methods can handle complex degradations, severe information loss may lead to semantic drift in the restored results. To address this issue, we propose RPL-UIE, a two-stage teacher--student framework for reliable prior learning. In the teacher stage, the network learns reliable and complementary spatial priors characterizing appearance and photometric properties from paired degraded and reference images. In the student stage, the network takes only degraded images as input and learns to emulate the teacher's prior extraction capability, thereby providing more reliable restoration guidance for the enhancement process without requiring reference images at inference. To reduce the prior-learning discrepancy between the teacher and student models, we further develop Residual Prior Refinement Diffusion (RPRD) and Frequency-Aware Prior Residual Calibration (FPRC). RPRD uses the coarse priors as anchors and progressively predicts the necessary corrections in the residual space. FPRC retains stable low-frequency residual components and selectively modulates high-frequency detail residuals, producing calibrated priors to support high-quality reconstruction. Experiments on multiple UIE benchmarks demonstrate competitive restoration performance. Downstream underwater object detection and instance segmentation experiments further demonstrate the improved utility of enhanced images for visual perception, while tests on real-world data captured by a remotely operated vehicle (ROV) support the robustness and practical applicability of RPL-UIE.

eess.IV↗

arXiv2026

RRTrack: Robust and Recoverable Object 6D Pose Tracking for Dynamic Scenes

Robust object 6D pose tracking is critical for robotic systems operating in dynamic and occluded scenes. Per-frame estimators are accurate but computationally expensive, while current trackers struggle with fast motion and complete occlusion due to their reliance on continuous visibility. To address these challenges, we present RRTrack, an efficient, recoverable object 6D pose tracker that enables robust tracking through fast motion and target disappearance--reappearance. RRTrack introduces a 2D--6D closed-loop tracking strategy that integrates memory-based video object segmentation (VOS) with 6D pose refinement. The 2D branch maintains target localization, and the 6D branch verifies geometric consistency before memory updates. In addition, a DINOv2-based dual-bank template matching module is developed to recover lost targets by jointly exploiting offline synthetic templates and online observation anchors while maintaining real-time efficiency. We also introduce a synthetic RGB-D benchmark comprising three robotic scenarios with fast motion and full occlusion. Experimental results on the synthetic benchmark demonstrate that RRTrack improves equal-subset mean ADD-S AR by 66.3\% and ADD-S AUC by 65.7\% over FoundationPose while achieving 55.2 FPS. Real-world experiments further validate the robustness of RRTrack under noisy sensing conditions. Project page: https://github.com/7kevin24/RRTrack

cs.CV↗

arXiv2026

Compressive sensing inspired self-supervised single-pixel imaging

Single-pixel imaging (SPI) is a promising imaging modality with distinctive advantages in strongly perturbed environments. Existing SPI methods lack physical sparsity constraints and overlook the integration of local and global features, leading to severe noise vulnerability, structural distortions and blurred details. To address these limitations, we propose SISTA-Net, a compressive sensing-inspired self-supervised method for single-pixel imaging. SISTA-Net unfolds the Iterative Shrinkage-Thresholding Algorithm (ISTA) into an interpretable network consisting of a data fidelity module and a proximal mapping module. The fidelity module adopts a hybrid CNN-Visual State Space Model (VSSM) architecture to integrate local and global feature modeling, enhancing reconstruction integrity and fidelity. We leverage deep nonlinear networks as adaptive sparse transforms combined with a learnable soft-thresholding operator to impose explicit physical sparsity in the latent domain, enabling noise suppression and robustness to interference even at extremely low sampling rates. Extensive experiments on multiple simulation scenarios demonstrate that SISTA-Net outperforms state-of-the-art methods by 2.6 dB in PSNR. Real-world far-field underwater tests yield a 3.4 dB average PSNR improvement, validating its robust anti-interference capability.

cs.CV↗

arXiv2026

Advancing Visual Reliability: Color-Accurate Underwater Image Enhancement for Real-Time Underwater Missions

Underwater image enhancement plays a crucial role in providing reliable visual information for underwater platforms, since strong absorption and scattering in water-related environments generally lead to image quality degradation. Existing high-performance methods often rely on complex architectures, which hinder deployment on underwater devices. Lightweight methods often sacrifice quality for speed and struggle to handle severely degraded underwater images. To address this limitation, we present a real-time underwater image enhancement framework with accurate color restoration. First, an Adaptive Weighted Channel Compensation module is introduced to achieve dynamic color recovery of the red and blue channels using the green channel as a reference anchor. Second, we design a Multi-branch Re-parameterized Dilated Convolution that employs multi-branch fusion during training and structural re-parameterization during inference, enabling large receptive field representation with low computational overhead. Finally, a Statistical Global Color Adjustment module is employed to optimize overall color performance based on statistical priors. Extensive experiments on eight datasets demonstrate that the proposed method achieves state-of-the-art performance across seven evaluation metrics. The model contains only 3,880 inference parameters and achieves an inference speed of 409 FPS. Our method improves the UCIQE score by 29.7% under diverse environmental conditions, and the deployment on ROV platforms and performance gains in downstream tasks further validate its superiority for real-time underwater missions.

cs.CV↗

arXiv2025

Mitigating Long-Tail Bias in HOI Detection via Adaptive Diversity Cache

Human-Object Interaction (HOI) detection is a fundamental task in computer vision, empowering machines to comprehend human-object relationships in diverse real-world scenarios. Recent advances in VLMs have significantly improved HOI detection by leveraging rich cross-modal representations. However, most existing VLM-based approaches rely heavily on additional training or prompt tuning, resulting in substantial computational overhead and limited scalability, particularly in long-tailed scenarios where rare interactions are severely underrepresented. In this paper, we propose the Adaptive Diversity Cache (ADC) module, a novel training-free and plug-and-play mechanism designed to mitigate long-tail bias in HOI detection. ADC constructs class-specific caches that accumulate high-confidence and diverse feature representations during inference. The method incorporates adaptive capacity allocation favoring rare categories and dynamic feature augmentation to enable robust prediction calibration without requiring additional training or fine-tuning. Extensive experiments on HICO-DET and V-COCO datasets show that ADC consistently improves existing HOI detectors, particularly enhancing rare category detection while preserving overall performance. These findings confirm the effectiveness of ADC as a training-free, plug-and-play solution for long-tail bias mitigation.

cs.CV↗

arXiv2025

A Generative Data Framework with Authentic Supervision for Underwater Image Restoration and Enhancement

Underwater image restoration and enhancement are crucial for correcting color distortion and restoring image details, thereby establishing a fundamental basis for subsequent underwater visual tasks. However, current deep learning methodologies in this area are frequently constrained by the scarcity of high-quality paired datasets. Since it is difficult to obtain pristine reference labels in underwater scenes, existing benchmarks often rely on manually selected results from enhancement algorithms, providing debatable reference images that lack globally consistent color and authentic supervision. This limits the model's capabilities in color restoration, image enhancement, and generalization. To overcome this limitation, we propose using in-air natural images as unambiguous reference targets and translating them into underwater-degraded versions, thereby constructing synthetic datasets that provide authentic supervision signals for model learning. Specifically, we establish a generative data framework based on unpaired image-to-image translation, producing a large-scale dataset that covers 6 representative underwater degradation types. The framework constructs synthetic datasets with precise ground-truth labels, which facilitate the learning of an accurate mapping from degraded underwater images to their pristine scene appearances. Extensive quantitative and qualitative experiments across 6 representative network architectures and 3 independent test sets show that models trained on our synthetic data achieve comparable or superior color restoration and generalization performance to those trained on existing benchmarks. This research provides a reliable and scalable data-driven solution for underwater image restoration and enhancement. The generated dataset is publicly available at: https://github.com/yftian2025/SynUIEDatasets.git.

cs.CV↗

arXiv2025

AudAgent: Automated Auditing of Privacy Policy Compliance in AI Agents

AI agents can autonomously perform tasks and, often without explicit user consent, collect or disclose users' sensitive local data, which raises serious privacy concerns. Although AI agents' privacy policies describe their intended data practices, there remains limited transparency and accountability about whether runtime behavior matches those policies. To bridge this gap, we present AudAgent, a tool that continuously monitors AI agents' data practices in real time and guards compliance with their stated privacy policies. AudAgent comprises four components for automated privacy auditing of AI agents. (i) Policy formalization: a novel cross-LLM voting mechanism that ensures high-confidence parsing of privacy policies into formal models. (ii) Runtime annotation: a lightweight Presidio-based analyzer that detects sensitive data and annotates data practices based on the AI agent's context and the formalized privacy policy model. (iii) Compliance auditing: ontology graphs and automata-based checking that link the privacy policy model with runtime annotations, enabling on-the-fly compliance verification. (iv) User interface: an infrastructure-independent implementation that visualizes the real-time execution trace of AI agents alongside detected privacy violations, providing user-friendly transparency and accountability. We evaluate AudAgent on AI agents built with mainstream frameworks, demonstrating its effectiveness in detecting and visualizing privacy policy violations. Using AudAgent, we further find that many privacy policies lack explicit safeguards for highly sensitive data such as SSNs, whose misuse violates legal requirements, and that many agents, including those powered by Claude, Gemini, and DeepSeek,do not refuse to process such data via third-party tools. AudAgent proactively blocks operations on such data, overriding the agents' original privacy policies and behavior.

cs.CR↗

arXiv2025

PrivAR: Client-Side Privacy Framework for Real-Time Location-Based Augmented Reality

Location-based augmented reality (LB-AR) applications, such as Pokemon Go, rely on sub-second GPS updates to deliver responsive and immersive user experiences. However, this high-frequency location reporting introduces serious privacy risks. Unlike traditional Location-Based Services (LBS), LB-AR demands real-time protection under strict latency and quality-of-service (QoS) constraints, while providing strong per-location and trajectory-level privacy guarantees. Existing privacy mechanisms struggle to satisfy these requirements: they either introduce prohibitive latency, significantly degrade application utility, or fail to defend against trajectory inference attacks. To address this challenge, we present PrivAR, the first client-side privacy framework for real-time LB-AR. PrivAR introduces two lightweight mechanisms: (i) Planar Staircase Mechanism (PSM), which uses a staircase-shaped distribution to generate noisy locations with strong per-location privacy, low expected distortion, and minimal computational overhead; and (ii) Planar Staircase Mechanism with Intermediate (PSM-I), an extension of PSM that generates a device-resident intermediate trajectory and selectively reuses previously perturbed outputs when insufficient drift is observed, thereby strengthening trace-level privacy while preserving high QoS. We provide theoretical analysis, extensive evaluation on two public mobility datasets and our GeoTrace dataset, and validate PrivAR in a Pokemon GO-style Android prototype. Results show that PrivAR improves AR QoS (game score) by up to 50% and increases attacker Bayes risk by up to 1.8x, while incurring only 0.06 ms of per-update overhead (less than 0.2% of end-to-end latency).

cs.CR↗

arXiv2025

Frequency Estimation of Correlated Multi-attribute Data under Local Differential Privacy

Large-scale data collection, from national censuses to IoT-enabled smart homes, routinely gathers dozens of attributes per individual. These multi-attribute datasets are crucial for analytics but pose significant privacy risks. Local Differential Privacy (LDP) is a powerful tool for protecting user privacy by allowing users to locally perturb their records before releasing them to an untrusted data aggregator. However, existing LDP mechanisms either split the privacy budget across all attributes or treat each attribute independently, thereby ignoring natural inter-attribute correlations. This leads to excessive noise and, consequently, significant utility loss, particularly in high-dimensional datasets. We introduce a two-phase LDP framework that overcomes these limitations by privately learning and exploiting inter-attribute dependencies. In Phase~I, a small subset of users applies a standard per-attribute LDP mechanism, enabling the aggregator to derive dependency information from the privatized data. In Phase~II, each remaining user perturbs a single randomly chosen attribute with the full privacy budget, while the unreported attributes are reconstructed using Phase~I statistics, incurring no additional privacy cost. As a concrete instantiation, we develop Correlated Randomized Response (Corr-RR), which employs correlation-aware probabilistic mappings to substantially improve estimation accuracy. We prove that Corr-RR satisfies $\epsilon$-LDP, and demonstrate through extensive experiments on synthetic and real-world datasets that it consistently outperforms state-of-the-art baselines, with the largest gains in high-dimensional and strongly correlated datasets.

cs.CR↗

arXiv2025

Bidirectional Prototype-Reward co-Evolution for Test-Time Adaptation of Vision-Language Models

Test-time adaptation (TTA) is crucial in maintaining performance of Vision Language Models (VLMs) when facing distribution shifts, particularly when the source data or target labels are inaccessible. Existing TTA methods predominantly leverage the output probability distribution of CLIP for feature evaluation, resulting in biases under domain shifts, which cause misclassified features due to text priors or incorrect textual associations. To address these issues, we propose \underline{B}idirectional Prototype-Reward co-Evolution (BPRE), a novel VLMs framework with TTA that integrates feature quality assessment with prototype evolution via a synergistic feedback loop. First, the Multi-dimensional Quality-aware Reward Module (MQRM) is designed to evaluate feature quality and guide prototype refinement precisely. The continuous refinement of prototype quality via Prototype-Reward Interactive Evolution (PRIE) enhances the computation more robust. Through this bidirectional interaction, the precision of rewards and prototype evolution mutually reinforce each other, forming a self-evolving feedback cycle. Extensive experiments conducted on 15 diverse recognition datasets demonstrate that our model consistently achieves superior performance compared to other SOTA methods, and advances VLM generalization capabilities through emphasizing comprehensive feature evaluation.

cs.CV↗

arXiv2025

Quantifying Classifier Utility under Local Differential Privacy

Local differential privacy (LDP) offers rigorous, quantifiable privacy guarantees for personal data by introducing perturbations at the data source. Understanding how these perturbations affect classifier utility is crucial for both designers and users. However, a general theoretical framework for quantifying this impact is lacking and also challenging, especially for complex or black-box classifiers. This paper presents a unified framework for theoretically quantifying classifier utility under LDP mechanisms. The key insight is that LDP perturbations are concentrated around the original data with a specific probability, allowing utility analysis to be reframed as robustness analysis within this concentrated region. Our framework thus connects the concentration properties of LDP mechanisms with the robustness of classifiers, treating LDP mechanisms as general distributional functions and classifiers as black boxes. This generality enables applicability to any LDP mechanism and classifier. A direct application of our utility quantification is guiding the selection of LDP mechanisms and privacy parameters for a given classifier. Notably, our analysis shows that a piecewise-based mechanism often yields better utility than alternatives in common scenarios. Beyond the core framework, we introduce two novel refinement techniques that further improve utility quantification. We then present case studies illustrating utility quantification for various combinations of LDP mechanisms and classifiers. Results demonstrate that our theoretical quantification closely matches empirical observations, particularly when classifiers operate in lower-dimensional input spaces.

cs.CR↗

arXiv2025

Locally Differentially Private Frequency Estimation via Joint Randomized Response

Local Differential Privacy (LDP) has been widely recognized as a powerful tool for providing a strong theoretical guarantee of data privacy to data contributors against an untrusted data collector. Under a typical LDP scheme, each data contributor independently randomly perturbs their data before submitting them to the data collector, which in turn infers valuable statistics about the original data from received perturbed data. Common to existing LDP mechanisms is an inherent trade-off between the level of privacy protection and data utility in the sense that strong data privacy often comes at the cost of reduced data utility. Frequency estimation based on Randomized Response (RR) is a fundamental building block of many LDP mechanisms. In this paper, we propose a novel Joint Randomized Response (JRR) mechanism based on correlated data perturbations to achieve locally differentially private frequency estimation. JRR divides data contributors into disjoint groups of two members and lets those in the same group jointly perturb their binary data to improve frequency-estimation accuracy and achieve the same level of data privacy by hiding the group membership information in contrast to the classical RR mechanism. Theoretical analysis and detailed simulation studies using both real and synthetic datasets show that JRR achieves the same level of data privacy as the classical RR mechanism while improving the frequency-estimation accuracy in the overwhelming majority of the cases by up to two orders of magnitude.

cs.CR↗

arXiv2025

Optimal Piecewise-based Mechanism for Collecting Bounded Numerical Data under Local Differential Privacy

Numerical data with bounded domains is a common data type in personal devices, such as wearable sensors. While the collection of such data is essential for third-party platforms, it raises significant privacy concerns. Local differential privacy (LDP) has been shown as a framework providing provable individual privacy, even when the third-party platform is untrusted. For numerical data with bounded domains, existing state-of-the-art LDP mechanisms are piecewise-based mechanisms, which are not optimal, leading to reduced data utility. This paper investigates the optimal design of piecewise-based mechanisms to maximize data utility under LDP. We demonstrate that existing piecewise-based mechanisms are heuristic forms of the $3$-piecewise mechanism, which is far from enough to study optimality. We generalize the $3$-piecewise mechanism to its most general form, i.e. $m$-piecewise mechanism with no pre-defined form of each piece. Under this form, we derive the closed-form optimal mechanism by combining analytical proofs and off-the-shelf optimization solvers. Next, we extend the generalized piecewise-based mechanism to the circular domain (along with the classical domain), defined on a cyclic range where the distance between the two endpoints is zero. By incorporating this property, we design the optimal mechanism for the circular domain, achieving significantly improved data utility compared with existing mechanisms. Our proposed mechanisms guarantee optimal data utility under LDP among all generalized piecewise-based mechanisms. We show that they also achieve optimal data utility in two common applications of LDP: distribution estimation and mean estimation. Theoretical analyses and experimental evaluations prove and validate the data utility advantages of our proposed mechanisms.

cs.CR↗

arXiv2025

ZeroBP: Learning Position-Aware Correspondence for Zero-shot 6D Pose Estimation in Bin-Picking

Bin-picking is a practical and challenging robotic manipulation task, where accurate 6D pose estimation plays a pivotal role. The workpieces in bin-picking are typically textureless and randomly stacked in a bin, which poses a significant challenge to 6D pose estimation. Existing solutions are typically learning-based methods, which require object-specific training. Their efficiency of practical deployment for novel workpieces is highly limited by data collection and model retraining. Zero-shot 6D pose estimation is a potential approach to address the issue of deployment efficiency. Nevertheless, existing zero-shot 6D pose estimation methods are designed to leverage feature matching to establish point-to-point correspondences for pose estimation, which is less effective for workpieces with textureless appearances and ambiguous local regions. In this paper, we propose ZeroBP, a zero-shot pose estimation framework designed specifically for the bin-picking task. ZeroBP learns Position-Aware Correspondence (PAC) between the scene instance and its CAD model, leveraging both local features and global positions to resolve the mismatch issue caused by ambiguous regions with similar shapes and appearances. Extensive experiments on the ROBI dataset demonstrate that ZeroBP outperforms state-of-the-art zero-shot pose estimation methods, achieving an improvement of 9.1% in average recall of correct poses.

cs.CV↗

arXiv2024

TraCS: Trajectory Collection in Continuous Space under Local Differential Privacy

Trajectory collection is essential for location-based services, yet it can reveal highly sensitive information about users, such as daily routines and activities, raising serious privacy concerns. Local Differential Privacy (LDP) offers strong privacy guarantees for users even when the data collector is untrusted. However, existing trajectory collection methods under LDP are largely confined to discrete location spaces, where the size of the location space affects both privacy guarantees and trajectory utility. Moreover, many real-world applications, such as flying trajectories or wearable-sensor traces, naturally operate in continuous spaces, making these discrete-space methods inadequate. This paper shifts the focus from discrete to continuous spaces for trajectory collection under LDP. We propose two methods: TraCS-D, which perturbs the direction and distance of locations, and TraCS-C, which perturbs the Cartesian coordinates of locations. Both methods are theoretically and experimentally analyzed for trajectory utility in continuous spaces. TraCS can also be applied to discrete spaces by rounding perturbed locations to any discrete space embedded in the continuous space. In this case, the privacy and utility guarantees of TraCS are independent of the number of locations in the space, and each perturbation requires only $\Theta(1)$ time complexity. Evaluation results on discrete location spaces validate the efficiency advantage and demonstrate that TraCS outperforms state-of-the-art methods with improved trajectory utility, particularly for large privacy parameters.

cs.CR↗

arXiv2023

Global-Local MAV Detection under Challenging Conditions based on Appearance and Motion

Visual detection of micro aerial vehicles (MAVs) has received increasing research attention in recent years due to its importance in many applications. However, the existing approaches based on either appearance or motion features of MAVs still face challenges when the background is complex, the MAV target is small, or the computation resource is limited. In this paper, we propose a global-local MAV detector that can fuse both motion and appearance features for MAV detection under challenging conditions. This detector first searches MAV target using a global detector and then switches to a local detector which works in an adaptive search region to enhance accuracy and efficiency. Additionally, a detector switcher is applied to coordinate the global and local detectors. A new dataset is created to train and verify the effectiveness of the proposed detector. This dataset contains more challenging scenarios that can occur in practice. Extensive experiments on three challenging datasets show that the proposed detector outperforms the state-of-the-art ones in terms of detection accuracy and computational efficiency. In particular, this detector can run with near real-time frame rate on NVIDIA Jetson NX Xavier, which demonstrates the usefulness of our approach for real-world applications. The dataset is available at https://github.com/WestlakeIntelligentRobotics/GLAD. In addition, A video summarizing this work is available at https://youtu.be/Tv473mAzHbU.

cs.CV↗

arXiv2023

Geo6D: Geometric Constraints Learning for 6D Pose Estimation

Numerous 6D pose estimation methods have been proposed that employ end-to-end regression to directly estimate the target pose parameters. Since the visible features of objects are implicitly influenced by their poses, the network allows inferring the pose by analyzing the differences in features in the visible region. However, due to the unpredictable and unrestricted range of pose variations, the implicitly learned visible feature-pose constraints are insufficiently covered by the training samples, making the network vulnerable to unseen object poses. To tackle these challenges, we proposed a novel geometric constraints learning approach called Geo6D for direct regression 6D pose estimation methods. It introduces a pose transformation formula expressed in relative offset representation, which is leveraged as geometric constraints to reconstruct the input and output targets of the network. These reconstructed data enable the network to estimate the pose based on explicit geometric constraints and relative offset representation mitigates the issue of the pose distribution gap. Extensive experimental results show that when equipped with Geo6D, the direct 6D methods achieve state-of-the-art performance on multiple datasets and demonstrate significant effectiveness, even with only 10% amount of data.

cs.CV↗

arXiv2023

Uni6Dv2: Noise Elimination for 6D Pose Estimation

Uni6D is the first 6D pose estimation approach to employ a unified backbone network to extract features from both RGB and depth images. We discover that the principal reasons of Uni6D performance limitations are Instance-Outside and Instance-Inside noise. Uni6D's simple pipeline design inherently introduces Instance-Outside noise from background pixels in the receptive field, while ignoring Instance-Inside noise in the input depth data. In this paper, we propose a two-step denoising approach for dealing with the aforementioned noise in Uni6D. To reduce noise from non-instance regions, an instance segmentation network is utilized in the first step to crop and mask the instance. A lightweight depth denoising module is proposed in the second step to calibrate the depth feature before feeding it into the pose regression network. Extensive experiments show that our Uni6Dv2 reliably and robustly eliminates noise, outperforming Uni6D without sacrificing too much inference efficiency. It also reduces the need for annotated real data that requires costly labeling.

cs.CV↗