arXiv subjects

Jiajia Liu

Publications and source records attributed to Jiajia Liu.

At least 37 records · Page 2Linked to original sources

arXiv2026

NTIRE 2026 Challenge on Bitstream-Corrupted Video Restoration: Methods and Results

This paper reports on the NTIRE 2026 Challenge on Bitstream-Corrupted Video Restoration (BSCVR). The challenge aims to advance research on recovering visually coherent videos from corrupted bitstreams, whose decoding often produces severe spatial-temporal artifacts and content distortion. Built upon recent progress in bitstream-corrupted video recovery, the challenge provides a common benchmark for evaluating restoration methods under realistic corruption settings. We describe the dataset, evaluation protocol, and participating methods, and summarize the final results and main technical trends. The challenge highlights the difficulty of this emerging task and provides useful insights for future research on robust video restoration under practical bitstream corruption.

cs.CV↗

arXiv2026

NTIRE 2026 The 3rd Restore Any Image Model (RAIM) Challenge: AI Flash Portrait (Track 3)

In this paper, we present a comprehensive overview of the NTIRE 2026 3rd Restore Any Image Model (RAIM) challenge, with a specific focus on Track 3: AI Flash Portrait. Despite significant advancements in deep learning for image restoration, existing models still encounter substantial challenges in real-world low-light portrait scenarios. Specifically, they struggle to achieve an optimal balance among noise suppression, detail preservation, and faithful illumination and color reproduction. To bridge this gap, this challenge aims to establish a novel benchmark for real-world low-light portrait restoration. We comprehensively evaluate the proposed algorithms utilizing a hybrid evaluation system that integrates objective quantitative metrics with rigorous subjective assessment protocols. For this competition, we provide a dataset containing 800 groups of real-captured low-light portrait data. Each group consists of a 1K-resolution low-light input image, a 1K ground truth (GT), and a 1K person mask. This challenge has garnered widespread attention from both academia and industry, attracting over 100 participating teams and receiving more than 3,000 valid submissions. This report details the motivation behind the challenge, the dataset construction process, the evaluation metrics, and the various phases of the competition. The released dataset and baseline code for this track are publicly available from the same \href{https://github.com/zsn1434/AI_Flash-BaseLine/tree/main}{GitHub repository}, and the official challenge webpage is hosted on \href{https://www.codabench.org/competitions/12885/}{CodaBench}.

cs.CV↗

arXiv2026

The Second Challenge on Cross-Domain Few-Shot Object Detection at NTIRE 2026: Methods and Results

Cross-domain few-shot object detection (CD-FSOD) remains a challenging problem for existing object detectors and few-shot learning approaches, particularly when generalizing across distinct domains. As part of NTIRE 2026, we hosted the second CD-FSOD Challenge to systematically evaluate and promote progress in detecting objects in unseen target domains under limited annotation conditions. The challenge received strong community interest, with 128 registered participants and a total of 696 submissions. Among them, 31 teams actively participated, and 19 teams submitted valid final results. Participants explored a wide range of strategies, introducing innovative methods that push the performance frontier under both open-source and closed-source tracks. This report presents a detailed overview of the NTIRE 2026 CD-FSOD Challenge, including a summary of the submitted approaches and an analysis of the final results across all participating teams. Challenge Codes: https://github.com/ohMargin/NTIRE2026_CDFSOD.

cs.CV↗

arXiv2026

NTIRE 2026 Challenge on Short-form UGC Video Restoration in the Wild with Generative Models: Datasets, Methods and Results

This paper presents an overview of the NTIRE 2026 Challenge on Short-form UGC Video Restoration in the Wild with Generative Models. This challenge utilizes a new short-form UGC (S-UGC) video restoration benchmark, termed KwaiVIR, which is contributed by USTC and Kuaishou Technology. It contains both synthetically distorted videos and real-world short-form UGC videos in the wild. For this edition, the released data include 200 synthetic training videos, 48 wild training videos, 11 validation videos, and 20 testing videos. The primary goal of this challenge is to establish a strong and practical benchmark for restoring short-form UGC videos under complex real-world degradations, especially in the emerging paradigm of generative-model-based S-UGC video restoration. This challenge has two tracks: (i) the primary track is a subjective track, where the evaluation is based on a user study; (ii) the second track is an objective track. These two tracks enable a comprehensive assessment of restoration quality. In total, 95 teams have registered for this competition. And 12 teams submitted valid final solutions and fact sheets for the testing phase. The submitted methods achieved strong performance on the KwaiVIR benchmark, demonstrating encouraging progress in short-form UGC video restoration in the wild.

cs.CV↗

arXiv2026

NTIRE 2026 The 3rd Restore Any Image Model (RAIM) Challenge: Multi-Exposure Image Fusion in Dynamic Scenes (Track 2)

This paper presents NTIRE 2026, the 3rd Restore Any Image Model (RAIM) challenge on multi-exposure image fusion in dynamic scenes. We introduce a benchmark that targets a practical yet difficult HDR imaging setting, where exposure bracketing must be fused under scene motion, illumination variation, and handheld camera jitter. The challenge data contains 100 training sequences with 7 exposure levels and 100 test sequences with 5 exposure levels, reflecting real-world scenarios that frequently cause misalignment and ghosting artefacts. We evaluate submissions with a leaderboard score derived from PSNR, SSIM, and LPIPS, while also considering perceptual quality, efficiency, and reproducibility during the final review. This track attracted 114 participating teams and received 987 submissions. The winning methods significantly improved the ability to remove artifacts from multi-exposure fusion and recover fine details. The dataset and the code of each team can be found at the repository: https://github.com/qulishen/RAIM-HDR.

cs.CV↗

arXiv2026

Rich-U-Net: A medical image segmentation model for fusing spatial depth features and capturing minute structural details

Medical image segmentation is of great significance in analysis of illness. The use of deep neural networks in medical image segmentation can help doctors extract regions of interest from complex medical images, thereby improving diagnostic accuracy and enabling better assessment of the condition to formulate treatment plans. However, most current medical image segmentation methods underperform in accurately extracting spatial information from medical images and mining potential complex structures and variations. In this article, we introduce the Rich-U-Net model, which effectively integrates both spatial and depth features. This fusion enhances the model's capability to detect fine structures and intricate details within complex medical images. Our multi-level and multi-dimensional feature fusion and optimization strategies enable our model to achieve fine structure localization and accurate segmentation results in medical image segmentation. Experiments on the ISIC2018, BUSI, GLAS, and CVC datasets show that Rich-U-Net surpasses other state-of-the-art models in Dice, IoU, and HD95 metrics.

eess.IV↗

arXiv2026

When Mamba Meets xLSTM: An Efficient and Precise Method with the xLSTM-VMUNet Model for Skin lesion Segmentation

Automatic melanoma segmentation is essential for early skin cancer detection, yet challenges arise from the heterogeneity of melanoma, as well as interfering factors like blurred boundaries, low contrast, and imaging artifacts. While numerous algorithms have been developed to address these issues, previous approaches have often overlooked the need to jointly capture spatial and sequential features within dermatological images. This limitation hampers segmentation accuracy, especially in cases with indistinct borders or structurally similar lesions. Additionally, previous models lacked both a global receptive field and high computational efficiency. In this work, we present the xLSTM-VMUNet Model, which jointly capture spatial and sequential features within dermatological images successfully. xLSTM-VMUNet can not only specialize in extracting spatial features from images, focusing on the structural characteristics of skin lesions, but also enhance contextual understanding, allowing more effective handling of complex medical image structures. Experiment results on the ISIC2018 dataset demonstrate that xLSTM-VMUNet outperforms VMUNet by 4.85% on DSC and 6.41% on IoU on the ISIC2017 dataset, by 1.25% on DSC and 2.07% on IoU on the ISIC2018 dataset, with faster convergence and consistently high segmentation performance. Our code is available at https://github.com/FangZhuoyi/XLSTM-VMUNet.

eess.IV↗

arXiv2026

No Need For Real Anomaly: MLLM Empowered Zero-Shot Video Anomaly Detection

The collection and detection of video anomaly data has long been a challenging problem due to its rare occurrence and spatio-temporal scarcity. Existing video anomaly detection (VAD) methods under perform in open-world scenarios. Key contributing factors include limited dataset diversity, and inadequate understanding of context-dependent anomalous semantics. To address these issues, i) we propose LAVIDA, an end-to-end zero-shot video anomaly detection framework. ii) LAVIDA employs an Anomaly Exposure Sampler that transforms segmented objects into pseudo-anomalies to enhance model adaptability to unseen anomaly categories. It further integrates a Multimodal Large Language Model (MLLM) to bolster semantic comprehension capabilities. Additionally, iii) we design a token compression approach based on reverse attention to handle the spatio-temporal scarcity of anomalous patterns and decrease computational cost. The training process is conducted solely on pseudo anomalies without any VAD data. Evaluations across four benchmark VAD datasets demonstrate that LAVIDA achieves SOTA performance in both frame-level and pixel-level anomaly detection under the zero-shot setting. Our code is available in https://github.com/VitaminCreed/LAVIDA.

cs.CV↗

arXiv2026

Inferring photospheric horizontal flows from multiple observations with SUVEL models

Photospheric horizontal velocity fields play essential roles in the formation and evolution of numerous solar activities. Various methods for estimating the horizontal velocity field have been proposed in the past. Aiming at the highest available (and future) spatial resolution (10 km/pixel) observations, a new method the Shallow U-net models (SUVEL) based on realistic numerical simulation and machine learning techniques was recently developed to track the photospheric horizontal velocity fields. Although SUVEL has been tested on numerical simulation data, its performance on solar observational data remained unclear. In this work, we apply SUVEL to the photospheric intensity observations from four ground-based solar telescopes (DKIST, GST, NVST, and SST) with the largest available apertures, and compare the results obtained from SUVEL with the Fourier local correlation tracking method (FLCT). Average correlation indices between granular regions and velocity fields inferred by SUVEL (FLCT) are 0.63, 0.81, 0.80, and 0.87 (0.00, 0.11, 0.16, and 0.10) for DKIST, GST, NVST, and SST observations. Higher correlation indices between the velocity fields tracked by SUVEL and granular patterns than FLCT reveal the superior performance of SUVEL, validating its reliability with respect to solar observational data.

astro-ph.SR↗

arXiv2025

A Machine Learning-Based Multimodal Framework for Wearable Sensor-Based Archery Action Recognition and Stress Estimation

In precision sports such as archery, athletes' performance depends on both biomechanical stability and psychological resilience. Traditional motion analysis systems are often expensive and intrusive, limiting their use in natural training environments. To address this limitation, we propose a machine learning-based multimodal framework that integrates wearable sensor data for simultaneous action recognition and stress estimation. Using a self-developed wrist-worn device equipped with an accelerometer and photoplethysmography (PPG) sensor, we collected synchronized motion and physiological data during real archery sessions. For motion recognition, we introduce a novel feature--Smoothed Differential Acceleration (SmoothDiff)--and employ a Long Short-Term Memory (LSTM) model to identify motion phases, achieving 96.8% accuracy and 95.9% F1-score. For stress estimation, we extract heart rate variability (HRV) features from PPG signals and apply a Multi-Layer Perceptron (MLP) classifier, achieving 80% accuracy in distinguishing high- and low-stress levels. The proposed framework demonstrates that integrating motion and physiological sensing can provide meaningful insights into athletes' technical and mental states. This approach offers a foundation for developing intelligent, real-time feedback systems for training optimization in archery and other precision sports.

cs.LG↗

arXiv2025

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer

Visual tokenization remains a core challenge in unifying visual understanding and generation within the autoregressive paradigm. Existing methods typically employ tokenizers in discrete latent spaces to align with the tokens from large language models, where the quantization errors can limit semantic expressiveness and degrade the capability of vision-language understanding. To address this, we introduce MingTok, a new family of visual tokenizers with a continuous latent space, for unified autoregressive generation and understanding. While understanding tasks favor discriminative high-dimensional features, generation tasks prefer compact low-level codes. Thus, to reconcile these competing demands, MingTok adopts a three-stage sequential architecture involving low-level encoding, semantic expansion, and visual reconstruction. Built on top of it, Ming-UniVision eliminates the need for task-specific visual representations, and unifies diverse vision-language tasks under a single autoregrsssive prediction paradigm. By formulating both understanding and generation as next-token prediction in a shared continuous space, it seamlessly supports multi-round, in-context tasks such as iterative understanding, generation and editing. Empirically, we find that using a unified continuous visual representation reconciles the competing requirements on the tokenizers by the understanding and generation tasks, thereby leading to state-of-the-art level performance across both domains. We hope our findings will facilitate unified visual tokenization in the continuous domain. Inference code and model weights are released to benefit community.

cs.CV↗

arXiv2025

HieraTok: Multi-Scale Visual Tokenizer Improves Image Reconstruction and Generation

In this work, we present HieraTok, a novel multi-scale Vision Transformer (ViT)-based tokenizer that overcomes the inherent limitation of modeling single-scale representations. This is realized through two key designs: (1) multi-scale downsampling applied to the token map generated by the tokenizer encoder, producing a sequence of multi-scale tokens, and (2) a scale-causal attention mechanism that enables the progressive flow of information from low-resolution global semantic features to high-resolution structural details. Coupling these designs, HieraTok achieves significant improvements in both image reconstruction and generation tasks. Under identical settings, the multi-scale visual tokenizer outperforms its single-scale counterpart by a 27.2\% improvement in rFID ($1.47 \rightarrow 1.07$). When integrated into downstream generation frameworks, it achieves a $1.38\times$ faster convergence rate and an 18.9\% boost in gFID ($16.4 \rightarrow 13.3$), which may be attributed to the smoother and more uniformly distributed latent space. Furthermore, by scaling up the tokenizer's training, we demonstrate its potential by a sota rFID of 0.45 and a gFID of 1.82 among ViT tokenizers. To the best of our knowledge, we are the first to introduce multi-scale ViT-based tokenizer in image reconstruction and image generation. We hope our findings and designs advance the ViT-based tokenizers in visual generation tasks.

cs.CV↗

arXiv2025

GUI-Shepherd: Reliable Process Reward and Verification for Long-Sequence GUI Tasks

Autonomous agents for long-sequence Graphical User Interface tasks are hindered by sparse rewards and the intractable credit assignment problem. To address these challenges, we introduce GUI-Shepherd, a Process Reward Model that provides dense, step-by-step feedback to guide agents. GUI-Shepherd is trained on a diverse large-scale data set of $52$k interactions that features human-annotated scores and GPT-4o generated rationales, enabling it to serve both as a reward provider for RL training and as a verifier for inference. As far as we know, we are the first to conduct a systematic study of process supervision in GUI agents, across diverse settings from online long-horizon tasks to offline single-step prediction. On the online AndroidWorld benchmark, GUI-Shepherd improves success rate by $7.7$ points via multi-turn online PPO, significantly outperforming Outcome Reward Model based competitors. When used as an inference verifier, it brings $5.1$ points improvements. The benefits generalize to the offline AndroidControl benchmark, with gains of $2.2$ points as a reward provider and $4.3$ points as a verifier. Collectively, our results establish that high-fidelity process supervision is critical for building more capable GUI agents and present a generalizable solution.

cs.AI↗

arXiv2025

Structuring GUI Elements through Vision Language Models: Towards Action Space Generation

Multimodal large language models (MLLMs) have emerged as pivotal tools in enhancing human-computer interaction. In this paper we focus on the application of MLLMs in the field of graphical user interface (GUI) elements structuring, where they assist in processing user instructions based on screen contents. Despite the promise of MLLMs, their performance in precisely generating UI element coordinates, a critical aspect of GUI understanding, is hindered by the nature of next-token prediction training. This challenge arises from the semantic void surrounding numerical UI coordinates in language representation spaces, necessitating a substantial and diverse dataset to bolster visual module capabilities. To address these limitations, we introduce an IoU-Augmented Maximum Likelihood (IAML) training paradigm. Specifically, our approach involves a novel pipeline for IoU-based coordinate sampling to augment the training data, which considers the proximity to ground truth coordinates. This data augmentation strategy is then employed to fine-tune MLLMs under the IAML paradigm, which is designed to mitigate the exposure bias problem inherent in traditional maximum likelihood estimation. Through extensive experiments, we demonstrate the superior performance of our IAML training approach over traditional training paradigms.

cs.CV↗

arXiv2025

A Multi-Layered Framework for Modeling Human Biology: From Basic AI Agents to a Full-Body AI Agent

We envision the Full-Body AI Agent as a comprehensive AI system designed to simulate, analyze, and optimize the dynamic processes of the human body across multiple biological levels. By integrating computational models, machine learning tools, and experimental platforms, this system aims to replicate and predict both physiological and pathological processes, ranging from molecules and cells to tissues, organs, and entire body systems. Central to the Full-Body AI Agent is its emphasis on integration and coordination across these biological levels, enabling analysis of how molecular changes influence cellular behaviors, tissue responses, organ function, and systemic outcomes. With a focus on biological functionality, the system is designed to advance the understanding of disease mechanisms, support the development of therapeutic interventions, and enhance personalized medicine. We propose two specialized implementations to demonstrate the utility of this framework: (1) the metastasis AI Agent, a multi-scale metastasis scoring system that characterizes tumor progression across the initiation, dissemination, and colonization phases by integrating molecular, cellular, and systemic signals; and (2) the drug AI Agent, a system-level drug development paradigm in which a drug AI-Agent dynamically guides preclinical evaluations, including organoids and chip-based models, by providing full-body physiological constraints. This approach enables the predictive modeling of long-term efficacy and toxicity beyond what localized models alone can achieve. These two agents illustrate the potential of Full-Body AI Agent to address complex biomedical challenges through multi-level integration and cross-scale reasoning.

q-bio.TO↗

arXiv2025

SARD: A YOLOv8-Based System for Solar Active Region Detection with SDO/HMI Magnetograms

Solar active regions are where sunspots are located and photospheric magnetic fluxes are concentrated, therefore being the sources of energetic eruptions in the solar atmosphere. The detection and statistics of solar active regions have been forefront topics in solar physics. In this study, we developed a solar active region detector (SARD) based on the advanced object detection model YOLOv8. First, we applied image processing techniques including thresholding and morphological operations to 6975 line-of-sight magnetograms from 2010 to 2019 at a cadence of 12~h, obtained by the Helioseismic and Magnetic Imager onboard the Solar Dynamic Observatory. With manual refinement, we labeled 26531 active regions in the dataset for further training and test with the detection model. Without any overlap between the training and test sets, the superior performance of SARD is demonstrated by an average precision rate as high as 94\%. We then performed a statistical analysis on the area and magnetic flux of the detected active regions, both of which yield log-normal distributions. This result sheds light on the underlying complexity and multi-scale nature of solar active regions.

astro-ph.SR↗

arXiv2025

BLAST: A Stealthy Backdoor Leverage Attack against Cooperative Multi-Agent Deep Reinforcement Learning based Systems

Recent studies have shown that cooperative multi-agent deep reinforcement learning (c-MADRL) is under the threat of backdoor attacks. Once a backdoor trigger is observed, it will perform malicious actions leading to failures or malicious goals. However, existing backdoor attacks suffer from several issues, e.g., instant trigger patterns lack stealthiness, the backdoor is trained or activated by an additional network, or all agents are backdoored. To this end, in this paper, we propose a novel backdoor leverage attack against c-MADRL, BLAST, which attacks the entire multi-agent team by embedding the backdoor only in a single agent. Firstly, we introduce adversary spatiotemporal behavior patterns as the backdoor trigger rather than manual-injected fixed visual patterns or instant status and control the period to perform malicious actions. This method can guarantee the stealthiness and practicality of BLAST. Secondly, we hack the original reward function of the backdoor agent via unilateral guidance to inject BLAST, so as to achieve the \textit{leverage attack effect} that can pry open the entire multi-agent system via a single backdoor agent. We evaluate our BLAST against 3 classic c-MADRL algorithms (VDN, QMIX, and MAPPO) in 2 popular c-MADRL environments (SMAC and Pursuit), and 2 existing defense mechanisms. The experimental results demonstrate that BLAST can achieve a high attack success rate while maintaining a low clean performance variance rate.

cs.AI↗

arXiv2025

M2-Reasoning: Empowering MLLMs with Unified General and Spatial Reasoning

Recent advancements in Multimodal Large Language Models (MLLMs), particularly through Reinforcement Learning with Verifiable Rewards (RLVR), have significantly enhanced their reasoning abilities. However, a critical gap persists: these models struggle with dynamic spatial interactions, a capability essential for real-world applications. To bridge this gap, we introduce M2-Reasoning-7B, a model designed to excel in both general and spatial reasoning. Our approach integrates two key innovations: (1) a novel data pipeline that generates 294.2K high-quality data samples (168K for cold-start fine-tuning and 126.2K for RLVR), which feature logically coherent reasoning trajectories and have undergone comprehensive assessment; and (2) a dynamic multi-task training strategy with step-wise optimization to mitigate conflicts between data, and task-specific rewards for delivering tailored incentive signals. This combination of curated data and advanced training allows M2-Reasoning-7B to set a new state-of-the-art (SOTA) across 8 benchmarks, showcasing superior performance in both general and spatial reasoning domains.

cs.AI↗