arXiv subjects

Luxi Zhao

Publications and source records attributed to Luxi Zhao.

18 recordsLinked to original sources

arXiv2026

DeepNC: A Fast GNN-based Pre-Verification Surrogate for TSN Configuration

Time-Sensitive Networking (TSN) is critical to deterministic communication in safety-critical domains, with formal verification such as Network Calculus (NC) serving as the cornerstone for schedulability guarantees. However, during automated configuration-space exploration, repeated schedulability analysis consumes over 90% of the total configuration time, becoming the primary bottleneck for large-scale TSN configurations. To address this challenge, we propose DeepNC, a novel pre-verification surrogate module that pioneers the structural fusion of NC principles into a Graph Neural Network (GNN) for TSN configuration-space exploration. Rather than replacing formal verification, DeepNC acts as a high-speed pre-verification filter, reserving computationally expensive formal verification only for promising candidates. Extensive evaluations demonstrate that DeepNC significantly improves worst-case delay prediction accuracy over state-of-the-art learning-based methods, increasing the average $R^2$ by 55.8% and reducing the average MAPE by 65.3%. More importantly, its high-fidelity regression substantially reduces the number of formal verification calls during configuration-space exploration by 93.25%, while accelerating NC-based verification by more than two orders of magnitude.

cs.NI↗

arXiv2026

Mitigating Content Shift and Hallucination in GenAI Image Editing via Structural Refinement

Generative AI (GenAI) image editors, such as Nano Banana, produce visually compelling results for retouching tasks, enabling non-experts to edit images through text prompts alone. However, the generative nature of these models often introduces spatial misalignment, texture distortion, and content hallucination, all of which are detrimental to downstream workflows that require pixel-level fidelity. We identify a problem setting we call "structure-preserving GenAI fusion" for black-box GenAI image retouching: retain the perceptual enhancements of a GenAI output while enforcing structural faithfulness to the original input image. To address this problem, we propose a post-processing framework that fuses an input image with its GenAI-enhanced counterpart by first establishing coarse spatial and photometric correspondences, then performing a fusion stage that transfers desired enhancements while suppressing hallucinated content. In the absence of direct prior work in this setting, we evaluate our framework against representative methods from photorealistic style transfer and image fusion. Our experiments demonstrate that our method better preserves aesthetic quality while maintaining pixel-level structural consistency and the input resolution.

cs.CV↗

arXiv2026

TSNBench: Benchmarking LLM Proficiency in Time-Sensitive Networking

We present TSNBench, the first benchmark for evaluating large language model (LLM) proficiency in Time-Sensitive Networking (TSN), a suite of IEEE 802.1 standards for deterministic communication with bounded latency in safety-critical domains such as autonomous vehicles, aviation, defense, and industrial automation. While LLMs have been extensively evaluated on general knowledge tasks, their capabilities in safety-critical networking domains remain largely unexplored. TSNBench comprises 939 expert-validated multiple-choice questions (MCQs) covering diverse TSN mechanisms, along with 100 open-ended Worst-Case Delay (WCD) computation tasks for Credit-Based Shaper (CBS) and Cyclic Queuing and Forwarding (CQF) across varying network topologies and traffic conditions. MCQ answers are validated by domain experts, and open-ended ground truth WCD values are computed using a verified Network Calculus (NC) solver for CBS and closed-form mathematical upper bounds for CQF. We evaluate 16 LLMs and find that although models achieve 67 to 95% accuracy on MCQs, they fail substantially on open-ended WCD computation. For CBS, only GPT-5 achieves a Mean Absolute Percentage Error (MAPE) of 36.2%, meaning its predicted WCD deviates by 36.2% of the actual TSN flow delay on average, while most models exceed 80%. For CQF, the best model achieves 41.8% MAPE, with most models clustering between 80% and 100%. Such errors are large relative to TSN latency budgets and can lead to violations of real-time constraints and unsafe configurations. TSNBench demonstrates that MCQ benchmarks may overestimate LLM capabilities in safety-critical networking domains.

cs.NI↗

arXiv2026

Addressing Image Authenticity When Cameras Use Generative AI

The ability of generative AI (GenAI) methods to photorealistically alter camera images has raised awareness about the authenticity of images shared online. Interestingly, images captured directly by our cameras are considered authentic and faithful. However, with the increasing integration of deep-learning modules into cameras' capture-time hardware -- namely, the image signal processor (ISP) -- there is now a potential for hallucinated content in images directly output by our cameras. Hallucinated capture-time image content is typically benign, such as enhanced edges or texture, but in certain operations, such as AI-based digital zoom or low-light image enhancement, hallucinations can potentially alter the semantics and interpretation of the image content. As a result, users may not realize that the content in their camera images is not authentic. This paper addresses this issue by enabling users to recover the 'unhallucinated' version of the camera image to avoid misinterpretation of the image content. Our approach works by optimizing an image-specific multi-layer perceptron (MLP) decoder together with a modality-specific encoder so that, given the camera image, we can recover the image before hallucinated content was added. The encoder and MLP are self-contained and can be applied post-capture to the image without requiring access to the camera ISP. Moreover, the encoder and MLP decoder require only 180 KB of storage and can be readily saved as metadata within standard image formats such as JPEG and HEIC.

cs.CV↗

arXiv2026

Innovation Discovery System for Networking Research

As networking systems become increasingly complex, achieving disruptive innovation grows more challenging. At the same time, recent progress in Large Language Models (LLMs) has shown strong potential for scientific hypothesis formation and idea generation. Nevertheless, applying LLMs effectively to networking research remains difficult for two main reasons: standalone LLMs tend to generate ideas by recombining existing solutions, and current open-source networking resources do not provide the structured, idea-level knowledge necessary for data-driven scientific discovery. To bridge this gap, we present SciNet, a research idea generation system specifically designed for networking. SciNet is built upon three key components: (1) constructing a networking-oriented scientific discovery dataset from top-tier networking conferences, (2) simulating the human idea discovery workflow through problem setting, inspiration retrieval, and idea generation, and (3) developing an idea evaluation method that jointly measures novelty and practicality. Experimental results show that \system consistently produces practical and novel networking research ideas across multiple LLM backbones, and outperforms standalone LLM-based generation in overall idea quality.

cs.NI↗

arXiv2025

Edit-aware RAW Reconstruction

Users frequently edit camera images post-capture to achieve their preferred photofinishing style. While editing in the RAW domain provides greater accuracy and flexibility, most edits are performed on the camera's display-referred output (e.g., 8-bit sRGB JPEG) since RAW images are rarely stored. Existing RAW reconstruction methods can recover RAW data from sRGB images, but these approaches are typically optimized for pixel-wise RAW reconstruction fidelity and tend to degrade under diverse rendering styles and editing operations. We introduce a plug-and-play, edit-aware loss function that can be integrated into any existing RAW reconstruction framework to make the recovered RAWs more robust to different rendering styles and edits. Our loss formulation incorporates a modular, differentiable image signal processor (ISP) that simulates realistic photofinishing pipelines with tunable parameters. During training, parameters for each ISP module are randomly sampled from carefully designed distributions that model practical variations in real camera processing. The loss is then computed in sRGB space between ground-truth and reconstructed RAWs rendered through this differentiable ISP. Incorporating our loss improves sRGB reconstruction quality by up to 1.5-2 dB PSNR across various editing conditions. Moreover, when applied to metadata-assisted RAW reconstruction methods, our approach enables fine-tuning for target edits, yielding further gains. Since photographic editing is the primary motivation for RAW reconstruction in consumer imaging, our simple yet effective loss function provides a general mechanism for enhancing edit fidelity and rendering flexibility across existing methods.

cs.CV↗

arXiv2025

Introducing Large Language Models into the Design Flow of Time-Sensitive Networking

The growing demand for real-time, safety-critical systems has significantly increased both the adoption and complexity of Time-Sensitive Networking (TSN). Configuring an optimized TSN network is highly challenging, requiring careful planning, design, analysis, verification, validation, and deployment. Large Language Models (LLMs) have recently demonstrated strong capabilities in solving complex tasks, positioning them as promising candidates for automating end-to-end TSN deployment and management, referred to as TSN orchestration. This paper outlines the steps involved in TSN orchestration and the associated challenges. To assess the capabilities of existing LLMs, we conduct an initial proof-of-concept case study focused on TSN tasks across multiple models. Building on these insights, we propose an LLM-assisted orchestration framework. Unlike prior research on LLMs in computer networks, which has concentrated on general configuration and management, TSN-specific orchestration has not yet been investigated. We present the building blocks for automating TSN using LLMs, describe the proposed pipeline, and analyze opportunities and limitations for real-world deployment. This work provides the first roadmap toward assessing the feasibility of LLM-assisted TSN orchestration.

cs.NI↗

arXiv2025

Improved Mapping Between Illuminations and Sensors for RAW Images

RAW images are unprocessed camera sensor output with sensor-specific RGB values based on the sensor's color filter spectral sensitivities. RAW images also incur strong color casts due to the sensor's response to the spectral properties of scene illumination. The sensor- and illumination-specific nature of RAW images makes it challenging to capture RAW datasets for deep learning methods, as scenes need to be captured for each sensor and under a wide range of illumination. Methods for illumination augmentation for a given sensor and the ability to map RAW images between sensors are important for reducing the burden of data capture. To explore this problem, we introduce the first-of-its-kind dataset comprising carefully captured scenes under a wide range of illumination. Specifically, we use a customized lightbox with tunable illumination spectra to capture several scenes with different cameras. Our illumination and sensor mapping dataset has 390 illuminations, four cameras, and 18 scenes. Using this dataset, we introduce a lightweight neural network approach for illumination and sensor mapping that outperforms competing methods. We demonstrate the utility of our approach on the downstream task of training a neural ISP. Link to project page: https://github.com/SamsungLabs/illum-sensor-mapping.

cs.CV↗

arXiv2025

Learning Camera-Agnostic White-Balance Preferences

The image signal processor (ISP) pipeline in modern cameras consists of several modules that transform raw sensor data into visually pleasing images in a display color space. Among these, the auto white balance (AWB) module is essential for compensating for scene illumination. However, commercial AWB systems often strive to compute aesthetic white-balance preferences rather than accurate neutral color correction. While learning-based methods have improved AWB accuracy, they typically struggle to generalize across different camera sensors -- an issue for smartphones with multiple cameras. Recent work has explored cross-camera AWB, but most methods remain focused on achieving neutral white balance. In contrast, this paper is the first to address aesthetic consistency by learning a post-illuminant-estimation mapping that transforms neutral illuminant corrections into aesthetically preferred corrections in a camera-agnostic space. Once trained, our mapping can be applied after any neutral AWB module to enable consistent and stylized color rendering across unseen cameras. Our proposed model is lightweight -- containing only $\sim$500 parameters -- and runs in just 0.024 milliseconds on a typical flagship mobile CPU. Evaluated on a dataset of 771 smartphone images from three different cameras, our method achieves state-of-the-art performance while remaining fully compatible with existing cross-camera AWB techniques, introducing minimal computational and memory overhead.

cs.CV↗

arXiv2025

Time-Aware Auto White Balance in Mobile Photography

Cameras rely on auto white balance (AWB) to correct undesirable color casts caused by scene illumination and the camera's spectral sensitivity. This is typically achieved using an illuminant estimator that determines the global color cast solely from the color information in the camera's raw sensor image. Mobile devices provide valuable additional metadata-such as capture timestamp and geolocation-that offers strong contextual clues to help narrow down the possible illumination solutions. This paper proposes a lightweight illuminant estimation method that incorporates such contextual metadata, along with additional capture information and image colors, into a compact model (~5K parameters), achieving promising results, matching or surpassing larger models. To validate our method, we introduce a dataset of 3,224 smartphone images with contextual metadata collected at various times of day and under diverse lighting conditions. The dataset includes ground-truth illuminant colors, determined using a color chart, and user-preferred illuminants validated through a user study, providing a comprehensive benchmark for AWB evaluation.

cs.CV↗

arXiv2025

Examining Joint Demosaicing and Denoising for Single-, Quad-, and Nona-Bayer Patterns

Camera sensors have color filters arranged in a mosaic layout, traditionally following the Bayer pattern. Demosaicing is a critical step camera hardware applies to obtain a full-channel RGB image. Many smartphones now have multiple sensors with different patterns, such as Quad-Bayer or Nona-Bayer. Most modern deep network-based models perform joint demosaicing and denoising with the strategy of training a separate network per pattern. Relying on individual models per pattern requires additional memory overhead and makes it challenging to switch quickly between cameras. In this work, we are interested in analyzing strategies for joint demosaicing and denoising for the three main mosaic layouts (1x1 Single-Bayer, 2x2 Quad-Bayer, and 3x3 Nona-Bayer). We found concatenating a three-channel mosaic embedding to the input image and training a unified demosaicing architecture yields results that outperform existing Quad-Bayer and Nona-Bayer models and are comparable to Single-Bayer models. Additionally, we describe a maskout strategy that enhances the model performance and facilitates dead pixel correction -- a step often overlooked by existing AI-based demosaicing models. As part of this effort, we captured a new demosaicing dataset of 638 RAW images that contain challenging scenes with patches annotated for training, validation, and testing. Code and data is available at https://github.com/SamsungLabs/unified-demosaicing.

eess.IV↗

arXiv2025

Efficient Adaptive Bandwidth Allocation for Deadline-Aware Online Admission Control in Time-Sensitive Networking

With the growing demand for dynamic real-time applications, online admission control for time-critical event-triggered (ET) traffic in Time-Sensitive Networking (TSN) has become a critical challenge. The main issue lies in dynamically allocating bandwidth with real-time guarantees in response to traffic changes while also meeting the requirements for rapid response, scalability, and high resource utilization in online scenarios. To address this challenge, we propose an online admission control method for ET traffic based on the TSN/ATS+CBS (asynchronous traffic shaper and credit-based shaper) architecture. This method provides a flexible framework for real-time guaranteed online admission control, supporting dynamic bandwidth allocation and reclamation at runtime without requiring global reconfiguration, thus improving scalability. Within this framework, we further integrate a novel strategy based on network calculus (NC) theory for efficient and high-utilization bandwidth reallocation. On the one hand, the strategy focuses on adaptively balancing residual bandwidth with deadline awareness to prevent bottleneck egress ports, thereby improving admission capacity. On the other hand, it employs a non-trivial analytical result to reduce the search space, accelerating the solving process. Experimental results from both large-scale synthetic and realistic test cases show that, compared to the state-of-the-art, our method achieves an average 56% increase in admitted flows and an average 92% reduction in admission time. Additionally, it postpones the occurrence of bottleneck egress ports and the first rejection of admission requests, thereby enhancing adaptability.

cs.NI↗

arXiv2025

Learning-Based Traffic Classification for Mixed-Critical Flows in Time-Sensitive Networking

Time-Sensitive Networking (TSN) supports multiple traffic types with diverse timing requirements, such as hard real-time (HRT), soft real-time (SRT), and Best Effort (BE) within a single network. To provide varying Quality of Service (QoS) for these traffic types, TSN incorporates different scheduling and shaping mechanisms. However, assigning traffic types to the proper scheduler or shaper, known as Traffic-Type Assignment (TTA), is a known NP-hard problem. Relying solely on domain expertise to make these design decisions can be inefficient, especially in complex network scenarios. In this paper, we present a proof-of-concept highlighting the advantages of a learning-based approach to the TTA problem. We formulate an optimization model for TTA in TSN and develop a Proximal Policy Optimization (PPO) based Deep Reinforcement Learning (DRL) model, called ``TTASelector'', to assign traffic types to TSN flows efficiently. Using synthetic and realistic test cases, our evaluation shows that TTASelector assigns a higher number of traffic types to HRT and SRT flows compared to the state-of-the-art Tabu Search-based metaheuristic method.

cs.NI↗

arXiv2024

CyclicSim: Comprehensive Evaluation of Cyclic Shapers in Time-Sensitive Networking

Cyclic Queuing and Forwarding (CQF) is a key Time-Sensitive Networking (TSN) shaping mechanism that ensures bounded latency using a simple gate control list (GCL). Recently, variants of CQF, including Cycle Specific Queuing and Forwarding (CSQF) and Multi Cyclic Queuing and Forwarding (MCQF), have emerged. While popular TSN mechanisms such as the Time-Aware Shaper (TAS), Asynchronous Traffic Shaper (ATS), Credit-Based Shaper (CBS), and Strict Priority (SP) have been extensively studied, cyclic shapers have not been thoroughly evaluated. This paper presents a comprehensive analysis of CQF, CSQF, and MCQF, providing insights into their performance. We quantify delays through simulations and quantitative analysis on both synthetic and realistic networks. For the first time, we introduce an open-source OMNeT++ and INET4.4 based framework capable of modeling all three cyclic shaper variants. Our tool facilitates the validation of new algorithms and serves as a benchmark for cyclic shapers. Our evaluations reveal that MCQF supports diverse timing requirements, whereas CSQF, with its additional queue, often results in larger delays and jitter for some TT flows compared to CQF. Additionally, CSQF does not demonstrate significant advantages in TSN networks where propagation delays are less critical than in wide-area networks (WANs).

cs.NI↗

arXiv2024

Quantifying the Impact of Frame Preemption on Combined TSN Shapers

Different scheduling mechanisms in Time Sensitive Networking (TSN) can be integrated together to design and support complex architectures with enhanced capabilities for mixed critical networks. Integrating Frame Preemption (FP) with Credit-Based Shaper (CBS) and Gate Control List (GCL) opens up different modes and configuration choices resulting in a complex evaluation of several possibilities and their impact on the Quality of Service (QoS). In this paper, we implement and quantify the integration of preemptive CBS with GCL by incorporating FP into the architecture. Our experiments show that the end-to-end delay of Audio Video Bridging (AVB) flows shaped by CBS reduces significantly (up to 40\%) when AVB flows are set to preemptable class. We further show that the jitter of Time Triggered (TT) traffic remains unaffected in "with Hold/Release" mode. Furthermore, we propose to introduce Guardband (GB) in the "without Hold/Release" to reduce the jitter of the TT flow. We compare all the different integration modes, starting with CBS with GCL, extending it further to FP. We evaluate all feasible combinations in both synthetic and realistic scenarios and offer recommendations for practical configuration methods.

cs.NI↗

arXiv2022

Quantitative Performance Comparison of Various Traffic Shapers in Time-Sensitive Networking

Owning to the sub-standards being developed by IEEE Time-Sensitive Networking (TSN) Task Group, the traditional IEEE 802.1 Ethernet is enhanced to support real-time dependable communications for future time- and safety-critical applications. Several sub-standards have been recently proposed that introduce various traffic shapers (e.g., Time-Aware Shaper (TAS), Asynchronous Traffic Shaper (ATS), Credit-Based Shaper (CBS), Strict Priority (SP)) for flow control mechanisms of queuing and scheduling, targeting different application requirements. These shapers can be used in isolation or in combination and there is limited work that analyzes, evaluates and compares their performance, which makes it challenging for end-users to choose the right combination for their applications. This paper aims at (i) quantitatively comparing various traffic shapers and their combinations, (ii) summarizing, classifying and extending the architectures of individual and combined traffic shapers and their Network calculus (NC)-based performance analysis methods and (iii) filling the gap in the timing analysis research on handling two novel hybrid architectures of combined traffic shapers, i.e., TAS+ATS+SP and TAS+ATS+CBS. A large number of experiments, using both synthetic and realistic test cases, are carried out for quantitative performance comparisons of various individual and combined traffic shapers, from the perspective of upper bounds of delay, backlog and jitter. To the best of our knowledge, we are the first to quantitatively compare the performance of the main traffic shapers in TSN. The paper aims at supporting the researchers and practitioners in the selection of suitable TSN sub-protocols for their use cases.

cs.NI↗

arXiv2021

Real-Time Guarantees for Critical Traffic in IEEE 802.1Qbv TSN Networks with Unscheduled and Unsynchronized End-Systems

Time-Sensitive Networking (TSN) aims to extend the IEEE 802.1Q Ethernet standard with real-time and time-aware capabilities. Each device's transmission of time-critical frames is done according to a so-called Gate Control List (GCL) schedule via the timed-gate mechanism described in IEEE 802.1Qbv. Most schedule generation mechanisms for TSN have a constraining assumption that both switches and end-systems in the network must have at least the TSN capabilities related to scheduled gates and time synchronization. However, many TSN networks use off-the-shelf end-systems, e.g., for providing sensor data, which are not scheduled and/or synchronized. In this paper, we propose a more flexible scheduling strategy that considers a worst-case delay analysis within the scheduling synthesis step, leveraging the solution's optimality to support TSN networks with unscheduled and unsynchronized end-systems while still being able to guarantee bounded latency for critical messages. Our method enables real-world systems that feature off-the-shelf microcontrollers and sensor nodes without TSN capabilities connected to state-of-the-art TSN networks to communicate critical messages in a real-time fashion. We evaluate our approach using both synthetic and real-world test cases, comparing it with existing scheduling mechanisms. Furthermore, we use OMNET++ to validate the generated GCL schedules.

cs.NI↗

arXiv2020

Latency Analysis of Multiple Classes of AVB Traffic in TSN with Standard Credit Behavior using Network Calculus

Time-Sensitive Networking (TSN) is a set of amendments that extend Ethernet to support distributed safety-critical and real-time applications in the industrial automation, aerospace and automotive areas. TSN integrates multiple traffic types and supports interactions in several combinations. In this paper we consider the configuration supporting Scheduled Traffic (ST) traffic scheduled based on Gate-Control-Lists (GCLs), Audio-Video-Bridging (AVB) traffic according to IEEE 802.1BA that has bounded latencies, and Best-Effort (BE) traffic, for which no guarantees are provided. The paper extends the timing analysis method to multiple AVB classes and proofs the credit bounds for multiple classes of AVB traffic, respectively under frozen and non-frozen behaviors of credit during guard band (GB). They are prerequisites for non-overflow credits of Credit-Based Shaper (CBS) and preventing starvation of AVB traffic. Moreover, this paper proposes an improved timing analysis method reducing the pessimism for the worst-case end-to-end delays of AVB traffic by considering the limitations from the physical link rate and the output of CBS. Finally, we evaluate the improved analysis method on both synthetic and real-world test cases, showing the significant reduction of pessimism on latency bounds compared to related work, and presenting the correctness validation compared with simulation results. We also compare the AVB latency bounds in the case of frozen and non-frozen credit during GB. Additionally, we evaluate the scalability of our method with variation of the load of ST flows and of the bandwidth reservation for AVB traffic.

cs.PF↗