arXiv subjects

Xinyuan Zheng

Publications and source records attributed to Xinyuan Zheng.

8 recordsLinked to original sources

arXiv2026

Instance-Guided Report Anchoring for Text-Free 3D Abnormality Segmentation in Chest CT

Accurate 3D abnormality segmentation in chest CT requires dense spatial supervision, but obtaining expert voxel-level labels is costly. Radiology reports, however, are routinely generated during clinical interpretation and contain instance-specific descriptions that can provide additional guidance without new dense annotation. Existing vision-language grounding methods typically require report-derived findings at inference, making localization dependent on paired text and limiting each forward pass to a queried finding. We propose Instance-Guided Report Anchoring (IGRA), a model-agnostic module that preserves the correspondence between each annotated abnormality instance and the report finding that describes it. IGRA pools each instance representation and anchors it to the corresponding finding embedding during training; all text-related components are discarded at inference. We further reformulate free-text grounding on ReXGroundingCT as multi-label volumetric segmentation by merging same-category instances, allowing all abnormality categories to be predicted in one image-only forward pass. IGRA improves Dice by 22.5% over the strongest image-only baseline (30.93 vs. 25.25) and is comparable to VoxTell on the single-finding subset (30.29 vs. 30.43). Applied unchanged to four standard 3D segmentation backbones, IGRA improves Dice and hit rate across all architectures. Zero-shot evaluation on LIDC-IDRI, PleThora, and a private in-house dataset further shows consistent gains over image-only baselines.

cs.CV↗

arXiv2026

Semantic Class Distribution Learning for Debiasing Semi-Supervised Medical Image Segmentation

Medical image segmentation is critical for computer-aided diagnosis. However, dense pixel-level annotation is time-consuming and costly, and medical datasets often exhibit severe class imbalance. Such an imbalance causes minority structures to be overwhelmed by dominant classes in feature representations, hindering the learning of discriminative features and making reliable segmentation particularly challenging. To address this, we propose the Semantic Class Distribution Learning (SCDL) framework, a plug-and-play module that mitigates supervision and representation biases by learning structured class-conditional feature distributions. SCDL integrates Class Distribution Bidirectional Alignment (CDBA) to align embeddings with learnable class proxies and leverages Semantic Anchor Constraints (SAC) to guide proxies using labeled data. Experiments on the Synapse and AMOS datasets demonstrate that SCDL largely improves segmentation performance across both overall and class-level metrics, with particularly notable gains for several low-frequency organs. Our anonymous code is released at https://anonymous.4open.science/r/SCDL.

cs.CV↗

arXiv2026

AGA3DNet: Anatomy-Guided Gaussian Priors with Multi-view xLSTM for 3D Brain MRI Subtype Classification

Accurate 3D brain MRI subtype classification benefits from both localized anatomical cues and long-range contextual reasoning. We present AGA3DNet, a report-grounded framework that incorporates brief anatomical phrases extracted from radiology reports as a soft anatomical prior channel and fuses it with a lightweight 3D CNN and multi-view xLSTM aggregation. Specifically, extracted anatomical phrases are mapped to atlas-defined regions and converted into smooth spatial priors using a signed-distance transform followed by Gaussian weighting, providing interpretable, anatomy-grounded guidance without requiring dense voxel annotations. We evaluate AGA3DNet on a retrospective institutional brain MRI cohort for abnormal subtype discrimination and compare against reproducible 3D classification baselines. AGA3DNet achieves improved overall balance across performance metrics and supports clinically interpretable localization through the prior channel. We discuss limitations related to single-cohort evaluation and the lack of large-scale public brain MRI datasets paired with radiology reports under broadly usable terms.

cs.CV↗

arXiv2025

Quantum simulation of interacting bosons with propagating waveguide photons

Optical networks composed of interconnected waveguides are a versatile platform to simulate bosonic physical phenomena. Significant work in the non-interacting regime has demonstrated the capabilities of this platform to simulate many exotic effects such as photon transport in the presence of gauge fields, dynamics of quantum walks, and topological transition and dissipation phenomena. However, the extension of these concepts to simulating interacting quantum many-body phenomena such as the Bose-Hubbard and the fractional quantum Hall (FQH) physics has remained elusive. In this work, we address this problem and demonstrate a framework for quantum many-body simulation as well as drive and dissipation in photonic waveguides. Specifically, we show that for waveguide photons, a tunable on-site interaction can be simulated using a photon-number-selective phase gate. We propose an implementation of such a phase gate based on a three-level-atom-mediated photon subtraction and addition. We apply this approach to bosonic lattice models and propose circuits that can accurately simulate the Bose-Hubbard and FQH Hamiltonian as benchmarking examples. Moreover, we show how to simulate the Lindbladian evolution with engineered dissipators such that the steady state of the Lindbadlian corresponds to the ground state of desired Hamiltonians. Our scheme extends the waveguide photonic simulation platform to the strongly interacting quantum many-body regime while retaining all of its crucial advantages, such as single-site addressability, Hamiltonian parameter controllability, and hardware efficiency.

quant-ph↗

arXiv2025

Denoising VAE as an Explainable Feature Reduction and Diagnostic Pipeline for Autism Based on Resting state fMRI

Autism spectrum disorders (ASDs) are developmental conditions characterized by restricted interests and difficulties in communication. The complexity of ASD has resulted in a deficiency of objective diagnostic biomarkers. Deep learning methods have gained recognition for addressing these challenges in neuroimaging analysis, but finding and interpreting such diagnostic biomarkers are still challenging computationally. Here, we propose a feature reduction pipeline using resting-state fMRI data. We used Craddock atlas and Power atlas to extract functional connectivity data from rs-fMRI, resulting in over 30 thousand features. By using a denoising variational autoencoder, our proposed pipeline further compresses the connectivity features into 5 latent Gaussian distributions, providing is a low-dimensional representation of the data to promote computational efficiency and interpretability. To test the method, we employed the extracted latent representations to classify ASD using traditional classifiers such as SVM on a large multi-site dataset. The 95% confidence interval for the prediction accuracy of SVM is [0.63, 0.76] after site harmonization using the extracted latent distributions. Without using DVAE for dimensionality reduction, the prediction accuracy is 0.70, which falls within the interval. The DVAE successfully encoded the diagnostic information from rs-fMRI data without sacrificing prediction performance. The runtime for training the DVAE and obtaining classification results from its extracted latent features was 7 times shorter compared to training classifiers directly on the raw data. Our findings suggest that the Power atlas provides more effective brain connectivity insights for diagnosing ASD than Craddock atlas. Additionally, we visualized the latent representations to gain insights into the brain networks contributing to the differences between ASD and neurotypical brains.

eess.IV↗

arXiv2025

Fed-NDIF: A Noise-Embedded Federated Diffusion Model For Low-Count Whole-Body PET Denoising

Low-count positron emission tomography (LCPET) imaging can reduce patients' exposure to radiation but often suffers from increased image noise and reduced lesion detectability, necessitating effective denoising techniques. Diffusion models have shown promise in LCPET denoising for recovering degraded image quality. However, training such models requires large and diverse datasets, which are challenging to obtain in the medical domain. To address data scarcity and privacy concerns, we combine diffusion models with federated learning -- a decentralized training approach where models are trained individually at different sites, and their parameters are aggregated on a central server over multiple iterations. The variation in scanner types and image noise levels within and across institutions poses additional challenges for federated learning in LCPET denoising. In this study, we propose a novel noise-embedded federated learning diffusion model (Fed-NDIF) to address these challenges, leveraging a multicenter dataset and varying count levels. Our approach incorporates liver normalized standard deviation (NSTD) noise embedding into a 2.5D diffusion model and utilizes the Federated Averaging (FedAvg) algorithm to aggregate locally trained models into a global model, which is subsequently fine-tuned on local datasets to optimize performance and obtain personalized models. Extensive validation on datasets from the University of Bern, Ruijin Hospital in Shanghai, and Yale-New Haven Hospital demonstrates the superior performance of our method in enhancing image quality and improving lesion quantification. The Fed-NDIF model shows significant improvements in PSNR, SSIM, and NMSE of the entire 3D volume, as well as enhanced lesion detectability and quantification, compared to local diffusion models and federated UNet-based models.

eess.IV↗

arXiv2024

Dynamic control of 2D non-Hermitian photonic corner states in synthetic dimensions

Non-Hermitian models describe the physics of ubiquitous open systems with gain and loss. One intriguing aspect of non-Hermitian models is their inherent topology that can produce intriguing boundary phenomena like resilient higher-order topological insulators (HOTIs) and non-Hermitian skin effects (NHSE). Recently, time-multiplexed lattices in synthetic dimensions have emerged as a versatile platform for the investigation of these effects free of geometric restrictions. Despite holding broad applications, studies of these effects have been limited to static cases so far, and full dynamical control over the non-Hermitian effects has remained elusive. Here, we demonstrate the emergence of topological non-Hermitian corner states with remarkable temporal controllability and robustness in a two-dimensional photonic synthetic time lattice. Specifically, we showcase various dynamic control mechanisms for light confinement and flow, including spatial mode tapering, sequential non-Hermiticity on-off switching, dynamical corner state relocation, and light steering. Moreover, we establish the corner state's robustness in the presence of intensity modulation randomness and quantitatively determine its breakdown regime. Our findings extend non-Hermitian and topological photonic effects into higher synthetic dimensions, offering remarkable flexibility and real-time control possibilities. This opens avenues for topological classification, quantum walk simulations of many-body dynamics, and robust Floquet engineering, free from the limitations of physical geometries.

physics.optics↗

arXiv2023

A strongly interacting photonic quantum walk using single atom beam splitters

Photonics provide an efficient way to implement quantum walks, the quantum analogue of classical random walk that demonstrates rich physics with potential applications. However, most photonic quantum walks do not involve photon interactions, which limits their potential to explore strongly-correlated many-body physics of light. We propose a strongly interacting discrete-time photonic quantum walk using a network of single atom beamsplitters. We calculate output statistics of the quantum walk for the case of two photons, which reveals the strongly-correlated transport of photons. Particularly, the walk can exhibit either boson-like or fermion-like statistics which is tunable by post-selecting the two-photon detection time interval. Also, the walk can sort different types of two-photon bound states into distinct pairs of output ports under certain conditions. These unique phenomena show that our quantum walk is an intriguing platform to explore strongly-correlated quantum many-body states of light. Finally, we propose an experimental realization based on time-multiplexed synthetic dimensions.

quant-ph↗