arXiv subjects

Fan Zhang

Publications and source records attributed to Fan Zhang.

At least 343 records · Page 19Linked to original sources

arXiv2025

Transcranial Photoacoustic Imaging for Human Intracranial Pressure Evaluation

Photoacoustic imaging (PAI), by combining high optical contrast with ultrasonic resolution, offers a promising noninvasive approach for dynamic monitoring of cerebral vasculature. However, transcranial PAI still faces significant challenges due to strong attenuation of both optical and acoustic signals by the skull. In this study, we propose a multi-wavelength photoacoustic tomography system and method for intracranial pressure (ICP) assessment, enabling visualization of cross-sectional structures of the middle cerebral artery (MCA) through the human temporal bone. By utilizing multi-wavelength excitation in the near-infrared-I (NIR-I) window, quantitative maps of blood oxygen saturation ($\mathbf{sO_2}$) are reconstructed, and the relationship between oxygenation dynamics and ICP variations is established. Experimental results demonstrate that the proposed system can successfully capture dynamic $\mathbf{sO_2}$ fluctuations in the MCA despite skull attenuation, revealing its characteristic responses to ICP changes. This work provides a high-precision, noninvasive imaging tool for early stroke diagnosis, cerebral vascular function assessment, and neurointerventional guidance, highlighting the clinical translational potential of PAI in neuroscience.

physics.med-ph↗

arXiv2025

Evaluation Agent: Efficient and Promptable Evaluation Framework for Visual Generative Models

Recent advancements in visual generative models have enabled high-quality image and video generation, opening diverse applications. However, evaluating these models often demands sampling hundreds or thousands of images or videos, making the process computationally expensive, especially for diffusion-based models with inherently slow sampling. Moreover, existing evaluation methods rely on rigid pipelines that overlook specific user needs and provide numerical results without clear explanations. In contrast, humans can quickly form impressions of a model's capabilities by observing only a few samples. To mimic this, we propose the Evaluation Agent framework, which employs human-like strategies for efficient, dynamic, multi-round evaluations using only a few samples per round, while offering detailed, user-tailored analyses. It offers four key advantages: 1) efficiency, 2) promptable evaluation tailored to diverse user needs, 3) explainability beyond single numerical scores, and 4) scalability across various models and tools. Experiments show that Evaluation Agent reduces evaluation time to 10% of traditional methods while delivering comparable results. The Evaluation Agent framework is fully open-sourced to advance research in visual generative models and their efficient evaluation.

cs.CV↗

arXiv2025

VBench-2.0: Advancing Video Generation Benchmark Suite for Intrinsic Faithfulness

Video generation has advanced significantly, evolving from producing unrealistic outputs to generating videos that appear visually convincing and temporally coherent. To evaluate these video generative models, benchmarks such as VBench have been developed to assess their faithfulness, measuring factors like per-frame aesthetics, temporal consistency, and basic prompt adherence. However, these aspects mainly represent superficial faithfulness, which focus on whether the video appears visually convincing rather than whether it adheres to real-world principles. While recent models perform increasingly well on these metrics, they still struggle to generate videos that are not just visually plausible but fundamentally realistic. To achieve real "world models" through video generation, the next frontier lies in intrinsic faithfulness to ensure that generated videos adhere to physical laws, commonsense reasoning, anatomical correctness, and compositional integrity. Achieving this level of realism is essential for applications such as AI-assisted filmmaking and simulated world modeling. To bridge this gap, we introduce VBench-2.0, a next-generation benchmark designed to automatically evaluate video generative models for their intrinsic faithfulness. VBench-2.0 assesses five key dimensions: Human Fidelity, Controllability, Creativity, Physics, and Commonsense, each further broken down into fine-grained capabilities. Tailored to individual dimensions, our evaluation framework integrates generalists such as SOTA VLMs and LLMs, and specialists, including anomaly detection methods proposed for video generation. We conduct extensive human annotations to ensure evaluation alignment with human judgment. By pushing beyond superficial faithfulness toward intrinsic faithfulness, VBench-2.0 aims to set a new standard for the next generation of video generative models in pursuit of intrinsic faithfulness.

cs.CV↗

arXiv2025

AdaptiveAE: An Adaptive Exposure Strategy for HDR Capturing in Dynamic Scenes

Mainstream high dynamic range imaging techniques typically rely on fusing multiple images captured with different exposure setups (shutter speed and ISO). A good balance between shutter speed and ISO is crucial for achieving high-quality HDR, as high ISO values introduce significant noise, while long shutter speeds can lead to noticeable motion blur. However, existing methods often overlook the complex interaction between shutter speed and ISO and fail to account for motion blur effects in dynamic scenes. In this work, we propose AdaptiveAE, a reinforcement learning-based method that optimizes the selection of shutter speed and ISO combinations to maximize HDR reconstruction quality in dynamic environments. AdaptiveAE integrates an image synthesis pipeline that incorporates motion blur and noise simulation into our training procedure, leveraging semantic information and exposure histograms. It can adaptively select optimal ISO and shutter speed sequences based on a user-defined exposure time budget, and find a better exposure schedule than traditional solutions. Experimental results across multiple datasets demonstrate that it achieves the state-of-the-art performance.

cs.CV↗

arXiv2025

Little Red reionization factories

In this note, we suggest the possibility that an elucidation of the nature of the numerous Little Red Dots (LRDs) at high redshifts, may facilitate the resolution of another concurrent cosmic puzzle, namely reionization. Specifically, it is hypothesized that intergalactic HI gas is compelled by the tidal field associated with a growing gravitational entropy, in the form of gas streams, into colliding at the LRD sites. The resulting shock heating and compression encourage starburst activities, which subsequently photo-ionize the gas into HII, that in turn shines into the rest-optical via bremsstrahlung radiation within the minimal-timescale-cutoff (i.e., red) regime, before escaping back into the intergalactic space, propelled by the newly fusion-injected energy.

astro-ph.GA↗

arXiv2025

SpikeSTAG: Spatial-Temporal Forecasting via GNN-SNN Collaboration

Spiking neural networks (SNNs), inspired by the spiking behavior of biological neurons, offer a distinctive approach for capturing the complexities of temporal data. However, their potential for spatial modeling in multivariate time-series forecasting remains largely unexplored. To bridge this gap, we introduce a brand new SNN architecture, which is among the first to seamlessly integrate graph structural learning with spike-based temporal processing for multivariate time-series forecasting. Specifically, we first embed time features and an adaptive matrix, eliminating the need for predefined graph structures. We then further learn sequence features through the Observation (OBS) Block. Building upon this, our Multi-Scale Spike Aggregation (MSSA) hierarchically aggregates neighborhood information through spiking SAGE layers, enabling multi-hop feature extraction while eliminating the need for floating-point operations. Finally, we propose a Dual-Path Spike Fusion (DSF) Block to integrate spatial graph features and temporal dynamics via a spike-gated mechanism, combining LSTM-processed sequences with spiking self-attention outputs, effectively improve the model accuracy of long sequence datasets. Experiments show that our model surpasses the state-of-the-art SNN-based iSpikformer on all datasets and outperforms traditional temporal models at long horizons, thereby establishing a new paradigm for efficient spatial-temporal modeling.

cs.LG↗

arXiv2025

Ear-Keeper: A Cross-Platform AI System for Rapid and Accurate Ear Disease Diagnosis

Early and accurate detection systems for ear diseases, powered by deep learning, are essential for preventing hearing impairment and improving population health. However, the limited diversity of existing otoendoscopy datasets and the poor balance between diagnostic accuracy, computational efficiency, and model size have hindered the translation of artificial intelligence (AI) algorithms into healthcare applications. In this study, we constructed a large-scale, multi-center otoendoscopy dataset covering eight common ear diseases and healthy cases. Building upon this resource, we developed Best-EarNet, an ultrafast and lightweight deep learning architecture integrating a novel Local-Global Spatial Feature Fusion Module with a multi-scale supervision strategy, enabling real-time and accurate classification of ear conditions. Leveraging transfer learning, Best-EarNet, with a model size of only 2.94 MB, achieved diagnostic accuracies of 95.23% on an internal test set (22,581 images) and 92.14% on an external test set (1,652 images), while requiring only 0.0125 seconds (80 frames per second) to process a single image on a standard CPU. Further subgroup analysis by gender and age showed consistently excellent performance of Best-EarNet across all demographic groups. To enhance clinical interpretability and user trust, we incorporated Grad-CAM-based visualization, highlighting the specific abnormal ear regions contributing to AI predictions. Most importantly, we developed Ear-Keeper, a cross-platform intelligent diagnosis system built upon Best-EarNet, deployable on smartphones, tablets, and personal computers. Ear-Keeper enables public users and healthcare providers to perform comprehensive real-time video-based ear canal screening, supporting early detection and timely intervention of ear diseases.

cs.CV↗

arXiv2025

Coarse-to-Fine Joint Registration of MR and Ultrasound Images via Imaging Style Transfer

We developed a pipeline for registering pre-surgery Magnetic Resonance (MR) images and post-resection Ultrasound (US) images. Our approach leverages unpaired style transfer using 3D CycleGAN to generate synthetic T1 images, thereby enhancing registration performance. Additionally, our registration process employs both affine and local deformable transformations for a coarse-to-fine registration. The results demonstrate that our approach improves the consistency between MR and US image pairs in most cases.

eess.IV↗

arXiv2025

Towards a General-Purpose Zero-Shot Synthetic Low-Light Image and Video Pipeline

Low-light conditions pose significant challenges for both human and machine annotation. This in turn has led to a lack of research into machine understanding for low-light images and (in particular) videos. A common approach is to apply annotations obtained from high quality datasets to synthetically created low light versions. In addition, these approaches are often limited through the use of unrealistic noise models. In this paper, we propose a new Degradation Estimation Network (DEN), which synthetically generates realistic standard RGB (sRGB) noise without the requirement for camera metadata. This is achieved by estimating the parameters of physics-informed noise distributions, trained in a self-supervised manner. This zero-shot approach allows our method to generate synthetic noisy content with a diverse range of realistic noise characteristics, unlike other methods which focus on recreating the noise characteristics of the training data. We evaluate our proposed synthetic pipeline using various methods trained on its synthetic data for typical low-light tasks including synthetic noise replication, video enhancement, and object detection, showing improvements of up to 24\% KLD, 21\% LPIPS, and 62\% AP$_{50-95}$, respectively.

cs.CV↗

arXiv2025

Neutralizing Token Aggregation via Information Augmentation for Efficient Test-Time Adaptation

Test-Time Adaptation (TTA) has emerged as an effective solution for adapting Vision Transformers (ViT) to distribution shifts without additional training data. However, existing TTA methods often incur substantial computational overhead, limiting their applicability in resource-constrained real-world scenarios. To reduce inference cost, plug-and-play token aggregation methods merge redundant tokens in ViTs to reduce total processed tokens. Albeit efficient, it suffers from significant performance degradation when directly integrated with existing TTA methods. We formalize this problem as Efficient Test-Time Adaptation (ETTA), seeking to preserve the adaptation capability of TTA while reducing inference latency. In this paper, we first provide a theoretical analysis from a novel mutual information perspective, showing that token aggregation inherently leads to information loss, which cannot be fully mitigated by conventional norm-tuning-based TTA methods. Guided by this insight, we propose to \textbf{N}eutralize Token \textbf{A}ggregation \textbf{v}ia \textbf{I}nformation \textbf{A}ugmentation (\textbf{NAVIA}). Specifically, we directly augment the [CLS] token embedding and incorporate adaptive biases into the [CLS] token in shallow layers of ViTs. We theoretically demonstrate that these augmentations, when optimized via entropy minimization, recover the information lost due to token aggregation. Extensive experiments across various out-of-distribution benchmarks demonstrate that NAVIA significantly outperforms state-of-the-art methods by over 2.5\%, while achieving an inference latency reduction of more than 20\%, effectively addressing the ETTA challenge.

cs.CV↗

arXiv2025

DDTracking: A Deep Generative Framework for Diffusion MRI Tractography with Streamline Local-Global Spatiotemporal Modeling

This paper presents DDTracking, a novel deep generative framework for diffusion MRI tractography that formulates streamline propagation as a conditional denoising diffusion process. In DDTracking, we introduce a dual-pathway encoding network that jointly models local spatial encoding (capturing fine-scale structural details at each streamline point) and global temporal dependencies (ensuring long-range consistency across the entire streamline). Furthermore, we design a conditional diffusion model module, which leverages the learned local and global embeddings to predict streamline propagation orientations for tractography in an end-to-end trainable manner. We conduct a comprehensive evaluation across diverse, independently acquired dMRI datasets, including both synthetic and clinical data. Experiments on two well-established benchmarks with ground truth (ISMRM Challenge and TractoInferno) demonstrate that DDTracking largely outperforms current state-of-the-art tractography methods. Furthermore, our results highlight DDTracking's strong generalizability across heterogeneous datasets, spanning varying health conditions, age groups, imaging protocols, and scanner types. Collectively, DDTracking offers anatomically plausible and robust tractography, presenting a scalable, adaptable, and end-to-end learnable solution for broad dMRI applications. Code is available at: https://github.com/yishengpoxiao/DDtracking.git

cs.CV↗

arXiv2025

HumanSAM: Classifying Human-centric Forgery Videos in Human Spatial, Appearance, and Motion Anomaly

Numerous synthesized videos from generative models, especially human-centric ones that simulate realistic human actions, pose significant threats to human information security and authenticity. While progress has been made in binary forgery video detection, the lack of fine-grained understanding of forgery types raises concerns regarding both reliability and interpretability, which are critical for real-world applications. To address this limitation, we propose HumanSAM, a new framework that builds upon the fundamental challenges of video generation models. Specifically, HumanSAM aims to classify human-centric forgeries into three distinct types of artifacts commonly observed in generated content: spatial, appearance, and motion anomaly. To better capture the features of geometry, semantics and spatiotemporal consistency, we propose to generate the human forgery representation by fusing two branches of video understanding and spatial depth. We also adopt a rank-based confidence enhancement strategy during the training process to learn more robust representation by introducing three prior scores. For training and evaluation, we construct the first public benchmark, the Human-centric Forgery Video (HFV) dataset, with all types of forgeries carefully annotated semi-automatically. In our experiments, HumanSAM yields promising results in comparison with state-of-the-art methods, both in binary and multi-class forgery classification.

cs.CV↗

arXiv2025

RecGPT Technical Report

Recommender systems are among the most impactful applications of artificial intelligence, serving as critical infrastructure connecting users, merchants, and platforms. However, most current industrial systems remain heavily reliant on historical co-occurrence patterns and log-fitting objectives, i.e., optimizing for past user interactions without explicitly modeling user intent. This log-fitting approach often leads to overfitting to narrow historical preferences, failing to capture users' evolving and latent interests. As a result, it reinforces filter bubbles and long-tail phenomena, ultimately harming user experience and threatening the sustainability of the whole recommendation ecosystem. To address these challenges, we rethink the overall design paradigm of recommender systems and propose RecGPT, a next-generation framework that places user intent at the center of the recommendation pipeline. By integrating large language models (LLMs) into key stages of user interest mining, item retrieval, and explanation generation, RecGPT transforms log-fitting recommendation into an intent-centric process. To effectively align general-purpose LLMs to the above domain-specific recommendation tasks at scale, RecGPT incorporates a multi-stage training paradigm, which integrates reasoning-enhanced pre-alignment and self-training evolution, guided by a Human-LLM cooperative judge system. Currently, RecGPT has been fully deployed on the Taobao App. Online experiments demonstrate that RecGPT achieves consistent performance gains across stakeholders: users benefit from increased content diversity and satisfaction, merchants and the platform gain greater exposure and conversions. These comprehensive improvement results across all stakeholders validates that LLM-driven, intent-centric design can foster a more sustainable and mutually beneficial recommendation ecosystem.

cs.IR↗

arXiv2025

iFANnpp: Nuclear Power Plant Digital Twin for Robots and Autonomous Intelligence

Robotics has gained attention in the nuclear industry due to its precision and ability to automate tasks. However, there is a critical need for advanced simulation and control methods to predict robot behavior and optimize plant performance, motivating the use of digital twins. Most existing digital twins do not offer a total design of a nuclear power plant. Moreover, they are designed for specific algorithms or tasks, making them unsuitable for broader research applications. In response, this work proposes a comprehensive nuclear power plant digital twin designed to improve real-time monitoring, operational efficiency, and predictive maintenance. A full nuclear power plant is modeled in Unreal Engine 5 and integrated with a high-fidelity Generic Pressurized Water Reactor Simulator to create a realistic model of a nuclear power plant and a real-time updated virtual environment. The virtual environment provides various features for researchers to easily test custom robot algorithms and frameworks.

cs.RO↗

arXiv2025

Fine-Grained Traffic Inference from Road to Lane via Spatio-Temporal Graph Node Generation

Fine-grained traffic management and prediction are fundamental to key applications such as autonomous driving, lane change guidance, and traffic signal control. However, obtaining lane-level traffic data has become a critical bottleneck for data-driven models due to limitations in the types and number of sensors and issues with the accuracy of tracking algorithms. To address this, we propose the Fine-grained Road Traffic Inference (FRTI) task, which aims to generate more detailed lane-level traffic information using limited road data, providing a more energy-efficient and cost-effective solution for precise traffic management. This task is abstracted as the first scene of the spatio-temporal graph node generation problem. We designed a two-stage framework--RoadDiff--to solve the FRTI task. solve the FRTI task. This framework leverages the Road-Lane Correlation Autoencoder-Decoder and the Lane Diffusion Module to fully utilize the limited spatio-temporal dependencies and distribution relationships of road data to accurately infer fine-grained lane traffic states. Based on existing research, we designed several baseline models with the potential to solve the FRTI task and conducted extensive experiments on six datasets representing different road conditions to validate the effectiveness of the RoadDiff model in addressing the FRTI task. The relevant datasets and code are available at https://github.com/ShuhaoLii/RoadDiff.

cs.AI↗

arXiv2025

Unusual Electron-Phonon Interactions in Highly Anisotropic Two-Dimensional $Ta_2$$Ni_3$$Te_5$

Electron-phonon interactions (EPIs) represent a fundamental cornerstone of condensed matter physics, commanding persistent attention due to their pivotal role in driving novel quantum phenomena within low-dimensional materials. Here, we unveil unusual anisotropic electron-phonon coupling behaviors in quasi-one-dimensional $Ta_2$$Ni_3$$Te_5$ nano-flakes through a powerful combination of angle-resolved polarized Raman spectroscopy and density functional perturbation theory (DFPT). High-resolution transmission electron microscopy and scanning tunneling microscopy directly visualize the pronounced quasi-one-dimensional atomic chains within the crystal structure, establishing a structural foundation for the observed anisotropic interactions. Our Raman investigations reveal remarkable polarization-dependent responses in $A_g$ phonon modes that deviate significantly from conventional behavior, which our theoretical analyses attribute to complex anisotropic electron-photon and electron-phonon interactions. Temperature-dependent Raman measurements further uncover an intriguing phonon decay mechanism involving both three- and four-phonon processes, with the latter showing significant contributions in some modes - a possible manifestation of strong anisotropic electron-phonon interactions. Beyond revealing $Ta_2$$Ni_3$$Te_5$ as an exceptional platform for exploring anisotropic EPIs, this work demonstrates that integrating angle-resolved polarized Raman spectroscopy with DFPT calculations offers a powerful methodology for investigating electron-phonon interactions in emerging low-dimensional quantum materials.

cond-mat.mtrl-sci↗

arXiv2025

A Novel Streamline-based diffusion MRI Tractography Registration Method with Probabilistic Keypoint Detection

Registration of diffusion MRI tractography is an essential step for analyzing group similarities and variations in the brain's white matter (WM). Streamline-based registration approaches can leverage the 3D geometric information of fiber pathways to enable spatial alignment after registration. Existing methods usually rely on the optimization of the spatial distances to identify the optimal transformation. However, such methods overlook point connectivity patterns within the streamline itself, limiting their ability to identify anatomical correspondences across tractography datasets. In this work, we propose a novel unsupervised approach using deep learning to perform streamline-based dMRI tractography registration. The overall idea is to identify corresponding keypoint pairs across subjects for spatial alignment of tractography datasets. We model tractography as point clouds to leverage the graph connectivity along streamlines. We propose a novel keypoint detection method for streamlines, framed as a probabilistic classification task to identify anatomically consistent correspondences across unstructured streamline sets. In the experiments, we compare several existing methods and show highly effective and efficient tractography registration performance.

cs.CV↗

arXiv2025

AFDM-Enabled Integrated Sensing and Communication: Theoretical Framework and Pilot Design

The integrated sensing and communication (ISAC) has been envisioned as one representative usage scenario of sixth-generation (6G) network. However, the unprecedented characteristics of 6G, especially the doubly dispersive channel, make classical ISAC waveforms rather challenging to guarantee a desirable performance level. The recently proposed affine frequency division multiplexing (AFDM) can attain full diversity even under doubly dispersive effects, thus becoming a competitive candidate for next-generation ISAC waveforms. Relevant investigations are still at an early stage, which involve only straightforward design lacking explicit theoretical analysis. This paper provides an in-depth investigation on AFDM waveform design for ISAC applications. Specifically, the closed-form Crámer-Rao bounds of target detection for AFDM are derived, followed by a demonstration on its merits over existing counterparts. Furthermore, we formulate the ambiguity function of the pilot-assisted AFDM waveform for the first time, revealing conditions for stable sensing performance. To further enhance both the communication and sensing performance of the AFDM waveform, we propose a novel pilot design by exploiting the characteristics of AFDM signals. The proposed design is analytically validated to be capable of optimizing the ambiguity function property and channel estimation accuracy simultaneously as well as overcoming the sensing and channel estimation range limitation originated from the pilot spacing. Numerical results have verified the superiority of the proposed pilot design in terms of dual-functional performance.

eess.SP↗