arXiv subjects

Kaiyuan Yang

Publications and source records attributed to Kaiyuan Yang.

At least 19 recordsLinked to original sources

arXiv2026

Intracranial Aneurysm Classification and Segmentation via Tri-Axial ROI and Multi-Task Learning

Intracranial aneurysms are often asymptomatic until rupture, which carries high mortality. Rupture risk assessment and treatment planning depend on both aneurysm morphology and anatomical location, yet existing automated methods remain limited to binary detection without fine-grained anatomical classification or multi-class segmentation. We present a multi-task framework that simultaneously performs multi-label classification, multi-class aneurysm segmentation, and multi-class vessel segmentation across 13 anatomical locations and four imaging modalities (CTA, MRA, T2, T1-post). Our two-stage approach combines a fast 2D tri-axial Region of Interest (ROI) extraction method with a 3D multi-task nnU-Net backbone. A dual-decoder design mitigates the extreme volume imbalance between aneurysm and vessel classes, while cross-attention pooling and modality-specific auxiliary heads improve feature learning across heterogeneous inputs. Our two-fold ensemble achieved 2nd place in the RSNA 2025 Intracranial Aneurysm Detection challenge. Code, model weights, a 3D Slicer plugin, and the corrected segmentation labels are publicly available.

cs.CV↗

arXiv2026

A Magnon-Based Electric Field Controlled Magnetoelectric Device for Energy-Efficient Logic-in-Memory

We demonstrate a non-volatile magnetoelectric magnonic memory (MEMM) that enables fully electrical write/read via direct magnon-driven sensing in an insulating antiferromagnet. A fabricated SrIrO3/La-BiFeO3/SrIrO3 trilayer exhibits sub-100 ps switching, a remnant polarization of 20 uC/cm2, and a readout voltage contrast close to 1mV between high and low-resistance states. To connect device physics to circuit behavior, we develop and experimentally validate a compact circuit model that captures spin Hall injection and spin transport. Simulations with optimized material parameters predict output voltages > 100mV, enabling cascading without external amplification. Using this framework, we design MEMM-based memory and logic blocks, including a 1T1R array, two inverter implementations (complementary two-device and single-device), and a three-input majority gate, and evaluate deep-pipelined operation. The model projects switching energies down to 1 aJ per operation and logic propagation delays of 30-60 ps, indicating MEMM as a promising platform for energy-constrained, high throughput computing.

cond-mat.mtrl-sci↗

arXiv2026

The TopCoW Challenge -- Topology-Aware Circle of Willis Segmentation for CT and MR Angiography

The Circle of Willis (CoW) is an important network of arteries connecting major circulations of the brain. Its vascular architecture is believed to influence the risk, severity, and outcome of serious neurovascular diseases. However, characterizing the highly variable CoW anatomy remains a manual and time-consuming expert task. The CoW is commonly imaged by two non-invasive angiographic imaging modalities, magnetic resonance angiography (MRA) and computed tomography angiography (CTA), yet few datasets with annotated CoW anatomy exist, and there have been no established benchmarks for comparing CoW segmentation algorithms. We organized the TopCoW benchmark challenge alongside the release of an annotated CoW dataset with 125 paired MRA and CTA scans from the same patients. Voxel-level annotations for 13 vessel components were created using virtual reality technology and verified by clinical experts. Participants submitted algorithms for CoW segmentation and variant classification, which we evaluated on internal and external test sets comprising 226 scans from over five centers. The benchmark includes voxel-level segmentation, CoW component detection, CoW variant classification, and two clinical application tasks. We received submissions from over 250 participants across six continents. Top-performing teams achieved over 90% Dice scores for CoW segmentation, over 80% F1 scores for detecting key vessel components, and over 70% balanced accuracy in CoW variant classification across nearly all test sets. The best algorithms also supported clinically relevant downstream tasks by accurately classifying fetal-type posterior cerebral arteries and localizing aneurysms in relation to CoW anatomy. This benchmark demonstrated the utility of CoW segmentation algorithms for some downstream clinical applications with explainability.

cs.CV↗

arXiv2026

Fleet: Few Shots Lead Effective AI-generated Image Detection

AI-generated image (AIGI) detection is undergoing a critical transition from laboratory benchmarks to open-world adversarial defense. The prevalent paradigm focuses on finding static feature spaces, assuming that some invariant artifacts learned from historical data can achieve universal zero-shot generalization. While achieving saturation on several AIGI benchmarks, this static hypothesis suffers a severe performance drop against rapidly evolving generators (e.g., SD3, Nano Banana Pro). To address these limitations, we propose that the field should expand beyond "static generalization" to a new paradigm of "dynamic adaptation". We introduce Fleet, a framework that pioneers a dynamic paradigm of continuous few-shot evolution, enabling rapid alignment with emerging generative threats. Fleet improves few-shot adaptation by replacing unconstrained feature updates with constrained routing correction, where avoidance routing redirects novel AI samples away from Non-AI-dominated routes within decoupled subspaces. To validate this, we present Treasure, a benchmark spanning 64 models and 360k images, featuring diverse architectures and 20 closed-source commercial engines. Experiments reveal that while static SOTA methods fail catastrophically on modern generators, Fleet restores performance from 20.4% to 73.1% with only 10-shot adaptation on "Doubao Seedream 4.0". Code and data are available at https://github.com/ICTMCG/Fleet .

cs.CV↗

arXiv2026

PowerGenie: Analytically-Guided Evolutionary Discovery of Superior Reconfigurable Power Converters

Discovering superior circuit topologies requires navigating an exponentially large design space-a challenge traditionally reserved for human experts. Existing AI methods either select from predefined templates or generate novel topologies at a limited scale without rigorous verification, leaving large-scale performance-driven discovery underexplored. We present PowerGenie, a framework for automated discovery of higher-performance reconfigurable power converters at scale. PowerGenie introduces: (1) an automated analytical framework that determines converter functionality and theoretical performance limits without component sizing or SPICE simulation, and (2) an evolutionary finetuning method that co-evolves a generative model with its training distribution through fitness selection and uniqueness verification. Unlike existing methods that suffer from mode collapse and overfitting, our approach achieves higher syntax validity, function validity, novelty rate, and figure-of-merit (FoM). PowerGenie discovers a novel 8-mode reconfigurable converter with 23% higher FoM than the best training topology. SPICE simulations confirm average absolute efficiency gains of 10% across 8 modes and up to 17% at a single mode. Code will be released upon publication.

cs.LG↗

arXiv2026

Multimodal Fusion of Skeleton Dynamics and Clinical Gait Features for Video-Based Cerebral Palsy Severity Assessment

Video-based gait analysis has become a promising approach for assessing motor impairment in children with cerebral palsy (CP). However, existing methods usually rely on either pose sequences or handcrafted gait features alone, making it difficult to simultaneously capture spatiotemporal motion patterns and clinically meaningful biomechanical information. To address this gap, we propose a multimodal fusion framework that integrates skeleton dynamics with contribution-guided clinically meaningful gait features. First, Grad-CAM analysis on a pre-trained ST-GCN backbone identified the most discriminative body keypoints, providing an interpretable basis for subsequent gait feature extraction. We then build a dual-stream architecture, with one stream modeling skeleton dynamics using ST-GCN and the other encoding gait geatures derived from the identified keypoints. By fusing the two streams through feature cross-attention improved four-level CP motor severity classification to 70.86%, outperforming the baseline by 5.6 percentage points. Overall, this work suggests that integrating skeleton dynamics with clinically meaningful gait descriptors can improve both prediction performance and biomechanical interpretability for video-based CP severity assessment.

eess.IV↗

arXiv2026

An Efficient Power Management Unit With Continuous MPPT and Energy Recycling for Wireless Millimetric Biomedical Implants

Biomedical implants offer transformative tools to improve medical outcomes. To realize minimally invasive implants with miniaturized volume and weight, wireless power transfer has been extensively studied to replace bulky batteries that dominate the volume of traditional implants and require surgical replacements. Ultra-sonic and magnetoelectric WPT modalities, which leverage low frequency acoustic electrical coupling for energy transduction, become viable solutions for mm-scale receivers. This work presents a fully integrated power management unit for ME WPT in millimetric implants. The PMU achieves load independent maximum power extraction and usage by continuously matching the impedance of the transducer, dynamically optimizing the power stage across varying input divided by load conditions, and reusing the storage energy to sustain the system when input power drops. Its parallel-input regulation and storing stages architecture prevent the cascading power loss. With the skewed-duty-cycle MPPT technique and regulation efficiency optimizer, the PMU achieves a peak MPPT efficiency of 98.5 percent and a peak system overall efficiency of 73.33 percent. Additionally, the PMU includes an adaptive high-voltage charging stage that charges the stimulation capacitor up to 12 V with an improved efficiency of 37.88 percent.

eess.SY↗

arXiv2025

DENTEX: Dental Enumeration and Tooth Pathosis Detection Benchmark for Panoramic X-ray

Panoramic X-rays are frequently used in dentistry for treatment planning, but their interpretation can be both time-consuming and prone to error. Artificial intelligence (AI) has the potential to aid in the analysis of these X-rays, thereby improving the accuracy of dental diagnoses and treatment plans. Nevertheless, designing automated algorithms for this purpose poses significant challenges, mainly due to the scarcity of annotated data and variations in anatomical structure. To address these issues, we organized the Dental Enumeration and Diagnosis on Panoramic X-rays Challenge (DENTEX) in association with the International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) in 2023. This challenge aims to promote the development of algorithms for multi-label detection of abnormal teeth, using three types of hierarchically annotated data: partially annotated quadrant data, partially annotated quadrant-enumeration data, and fully annotated quadrant-enumeration-diagnosis data, inclusive of four different diagnoses. In this paper, we present a comprehensive analysis of the methods and results from the challenge. Our findings reveal that top performers succeeded through diverse, specialized strategies, from segmentation-guided pipelines to highly-engineered single-stage detectors, using advanced Transformer and diffusion models. These strategies significantly outperformed traditional approaches, particularly for the challenging tasks of tooth enumeration and subtle disease classification. By dissecting the architectural choices that drove success, this paper provides key insights for future development of AI-powered tools that can offer more precise and efficient diagnosis and treatment planning in dentistry. The evaluation code and datasets can be accessed at https://github.com/ibrahimethemhamamci/DENTEX

cs.CV↗

arXiv2025

Compressive Sensing Photoacoustic Imaging Receiver with Matrix-Vector-Multiplication SAR ADC

Wearable photoacoustic imaging devices hold great promise for continuous health monitoring and point-of-care diagnostics. However, the large data volume generated by high-density transducer arrays presents a major challenge for realizing compact and power-efficient wearable systems. This paper presents a photoacoustic imaging receiver (RX) that embeds compressive sensing directly into the hardware to address this bottleneck. The RX integrates 16 AFEs and four matrix-vector-multiplication (MVM) SAR ADCs that perform energy- and area-efficient analog-domain compression. The architecture achieves a 4-8x reduction in output data rate while preserving low-loss full-array information. The MVM SAR ADC executes passive and accurate MVM using user-defined programmable ternary weights. Two signal reconstruction methods are implemented: (1) an optimization approach using the fast iterative shrinkage-thresholding algorithm, and (2) a learning-based approach employing implicit neural representation. Fabricated in 65 nm CMOS, the chip achieves an ADC's SNDR of 57.5 dB at 20.41 MS/s, with an AFE input-referred noise of 3.5 nV/sqrt(Hz). MVM linearity measurements show R^2 > 0.999 across a wide range of weights and input amplitudes. The system is validated through phantom imaging experiments, demonstrating high-fidelity image reconstruction under up to 8x compression. The RX consumes 5.83 mW/channel and supports a general ternary-weighted measurement matrix, offering a compelling solution for next-generation miniaturized, wearable PA imaging systems.

eess.IV↗

arXiv2025

Circle of Willis Centerline Graphs: A Dataset and Baseline Algorithm

The Circle of Willis (CoW) is a critical network of arteries in the brain, often implicated in cerebrovascular pathologies. Voxel-level segmentation is an important first step toward an automated CoW assessment, but a full quantitative analysis requires centerline representations. However, conventional skeletonization techniques often struggle to extract reliable centerlines due to the CoW's complex geometry, and publicly available centerline datasets remain scarce. To address these challenges, we used a thinning-based skeletonization algorithm to extract and curate centerline graphs and morphometric features from the TopCoW dataset, which includes 200 stroke patients, each imaged with MRA and CTA. The curated graphs were used to develop a baseline algorithm for centerline and feature extraction, combining U-Net-based skeletonization with A* graph connection. Performance was evaluated on a held-out test set, focusing on anatomical accuracy and feature robustness. Further, we used the extracted features to predict the frequency of fetal PCA variants, confirm theoretical bifurcation optimality relations, and detect subtle modality differences. The baseline algorithm consistently reconstructed graph topology with high accuracy (F1 = 1), and the average Euclidean node distance between reference and predicted graphs was below one voxel. Features such as segment radius, length, and bifurcation ratios showed strong robustness, with median relative errors below 5% and Pearson correlations above 0.95. Our results demonstrate the utility of learning-based skeletonization combined with graph connection for anatomically plausible centerline extraction. We emphasize the importance of going beyond simple voxel-based measures by evaluating anatomical accuracy and feature robustness. The dataset and baseline algorithm have been released to support further method development and clinical research.

cs.CV↗

arXiv2025

ISLES'24: Final Infarct Prediction with Multimodal Imaging and Clinical Data. Where Do We Stand?

Accurate estimation of brain infarction (i.e., irreversibly damaged tissue) is critical for guiding treatment decisions in acute ischemic stroke. Reliable infarct prediction informs key clinical interventions, including the need for patient transfer to comprehensive stroke centers, the potential benefit of additional reperfusion attempts during mechanical thrombectomy, decisions regarding secondary neuroprotective treatments, and ultimately, prognosis of clinical outcomes. This work introduces the Ischemic Stroke Lesion Segmentation (ISLES) 2024 challenge, which focuses on the prediction of final infarct volumes from pre-interventional acute stroke imaging and clinical data. ISLES24 provides a comprehensive, multimodal setting where participants can leverage all clinically and practically available data, including full acute CT imaging, sub-acute follow-up MRI, and structured clinical information, across a train set of 150 cases. On the hidden test set of 98 cases, the top-performing model, a multimodal nnU-Net-based architecture, achieved a Dice score of 0.285 (+/- 0.213) and an absolute volume difference of 21.2 (+/- 37.2) mL, underlining the significant challenges posed by this task and the need for further advances in multimodal learning. This work makes two primary contributions: first, we establish a standardized, clinically realistic benchmark for post-treatment infarct prediction, enabling systematic evaluation of multimodal algorithmic strategies on a longitudinal stroke dataset; second, we analyze current methodological limitations and outline key research directions to guide the development of next-generation infarct prediction models.

eess.IV↗

arXiv2025

ISLES'24 -- A Real-World Longitudinal Multimodal Stroke Dataset

Stroke remains a leading cause of global morbidity and mortality, imposing a heavy socioeconomic burden. Advances in endovascular reperfusion therapy and CT and MR imaging for treatment guidance have significantly improved patient outcomes. Developing machine learning algorithms that can create accurate models of brain function from stroke images for tasks like lesion identification and tissue survival prediction requires large, diverse, and well annotated public datasets. While several high-quality image datasets in stroke exist, they include only single time point data. Data over different time points are essential to accurately identify lesions and predict prognosis. Here, we provide comprehensive longitudinal stroke data, including (sub-)acute CT imaging with angiography and perfusion, follow-up MRI after 2-9 days, and acute and longitudinal clinical data up to a three-month outcome. The dataset also includes vessel occlusion masks from acute CT angiography and delineated infarction masks in follow-up MRI. This multicenter dataset consists of 245 cases and is a solid basis for developing powerful machine-learning algorithms to facilitate clinical decision-making.

cs.CV↗

arXiv2025

Atmospheric Turbulence-Resilient Long-Range Fourier Ptychography

While Fourier ptychography (FP) offers super-resolution for macroscopic imaging, its real-world application is severely hampered by atmospheric turbulence, a challenge largely unaddressed in existing macroscopic FP research operating under idealized conditions. This work establishes, to our knowledge, the first comprehensive computational framework specifically designed for turbulence mitigation in long-range FP, termed Turbulence-Mitigated FP (TMFP). Rather than correcting pupil errors, an image degradation model is developed alongside a reconstruction pipeline inspired by speckle interferometry. By taking multiple short-exposure randomly-distorted measurements and exploiting their statistical properties, the diffraction-limited sub-aperture images can be recovered for further FP reconstruction. Numerical simulations and experimental validations under optical turbulence demonstrate the method's robustness, resolution enhancement, and practicality in adverse conditions, paving the way for the reliable deployment of high-resolution macroscopic FP in real-world scenarios.

physics.optics↗

arXiv2025

Parametric shape models for vessels learned from segmentations via differentiable voxelization

Vessels are complex structures in the body that have been studied extensively in multiple representations. While voxelization is the most common of them, meshes and parametric models are critical in various applications due to their desirable properties. However, these representations are typically extracted through segmentations and used disjointly from each other. We propose a framework that joins the three representations under differentiable transformations. By leveraging differentiable voxelization, we automatically extract a parametric shape model of the vessels through shape-to-segmentation fitting, where we learn shape parameters from segmentations without the explicit need for ground-truth shape parameters. The vessel is parametrized as centerlines and radii using cubic B-splines, ensuring smoothness and continuity by construction. Meshes are differentiably extracted from the learned shape parameters, resulting in high-fidelity meshes that can be manipulated post-fit. Our method can accurately capture the geometry of complex vessels, as demonstrated by the volumetric fits in experiments on aortas, aneurysms, and brain vessels.

cs.CV↗

arXiv2025

FlexDuo: A Pluggable System for Enabling Full-Duplex Capabilities in Speech Dialogue Systems

Full-Duplex Speech Dialogue Systems (Full-Duplex SDS) have significantly enhanced the naturalness of human-machine interaction by enabling real-time bidirectional communication. However, existing approaches face challenges such as difficulties in independent module optimization and contextual noise interference due to highly coupled architectural designs and oversimplified binary state modeling. This paper proposes FlexDuo, a flexible full-duplex control module that decouples duplex control from spoken dialogue systems through a plug-and-play architectural design. Furthermore, inspired by human information-filtering mechanisms in conversations, we introduce an explicit Idle state. On one hand, the Idle state filters redundant noise and irrelevant audio to enhance dialogue quality. On the other hand, it establishes a semantic integrity-based buffering mechanism, reducing the risk of mutual interruptions while ensuring accurate response transitions. Experimental results on the Fisher corpus demonstrate that FlexDuo reduces the false interruption rate by 24.9% and improves response accuracy by 7.6% compared to integrated full-duplex dialogue system baselines. It also outperforms voice activity detection (VAD) controlled baseline systems in both Chinese and English dialogue quality. The proposed modular architecture and state-based dialogue model provide a novel technical pathway for building flexible and efficient duplex dialogue systems.

cs.CL↗

arXiv2025

NLS: Natural-Level Synthesis for Hardware Implementation Through GenAI

This paper introduces Natural-Level Synthesis, an innovative approach for generating hardware using generative artificial intelligence on both the system level and component-level. NLS bridges a gap in current hardware development processes, where algorithm and application engineers' involvement typically ends at the requirements stage. With NLS, engineers can participate more deeply in the development, synthesis, and test stages by using Gen-AI models to convert natural language descriptions directly into Hardware Description Language code. This approach not only streamlines hardware development but also improves accessibility, fostering a collaborative workflow between hardware and algorithm engineers. We developed the NLS tool to facilitate natural language-driven HDL synthesis, enabling rapid generation of system-level HDL designs while significantly reducing development complexity. Evaluated through case studies and benchmarks using Performance, Power, and Area metrics, NLS shows its potential to enhance resource efficiency in hardware development. This work provides a extensible, efficient solution for hardware synthesis and establishes a Visual Studio Code Extension to assess Gen-AI-driven HDL generation and system integration, laying a foundation for future AI-enhanced and AI-in-the-loop Electronic Design Automation tools.

cs.AR↗

arXiv2025

DeepSeek Powered Solid Dosage Formulation Design and Development

Pharmaceutical process design and development for generic, innovative, or personalized drugs have always been a time-consuming, costly, rigorous process, that involves multi-stage evaluation for better quality control and assurance. Large language models (LLMs), a type of generative artificial intelligence system, can augment laboratory research in the pharmaceutical engineering process by helping scientists to extract knowledge from literature, design parameters, and collect and interpret experimental data ultimately accelerating scientific discovery. LLMs with prompt engineering technologies change the researchers thinking protocol from traditional empirical knowledge to streamlined thinking that connects the performance and structured parameters together. In this work, we investigate and evaluate how prompt engineering technologies can enhance the drug design process from different strategies such as zero-shot, few-shot, chain-of-thought, etc. The dissolution profile for specific drugs is predicted and suggested from the LLMs model. Furthermore, the fundamental physical properties such as PSD, aspect ratio, and specific surface area could be inversely designed from the LLMs model. Finally, all the results are evaluated and validated by real-world cases to prove the reliability of prompt engineering techniques. Initial evaluations show an MSE of 23.61 and R2 of 0.97 in zero-shot, an MSE of 114.89 and R2 of 0.90 in zero-shot-CoT, an MSE of 57.0 and R2 of 0.92 in few-shot, a MSE of 22.56 and R2 of 0.97 in few-shot-CoT and a MSE of 10.56 and R2 of 0.99 with the involvement of RAG. This work breaks down any barriers in developing a systematic framework where LLMs assist in formulation design, process control, and decision-making. Finally, we conclude the work by discussing open challenges and future research directions in pharmaceutical processes.

cs.ET↗

arXiv2025

Design-Agnostic Distributed Timing Fault Injection Monitor With End-to-End Design Automation

Fault injection attacks induce hardware failures in circuits and exploit these faults to compromise the security of the system. It has been demonstrated that FIAs can bypass system security mechanisms, cause faulty outputs, and gain access to secret information. Certain types of FIAs can be mounted with little effort by tampering with clock signals and or the chip operating conditions. To mitigate such low cost, yet powerful attacks, we propose a fully synthesizable and distributable in situ fault injection monitor that employs a delay locked loop to track the pulsewidth of the clock. We further develop a fully automated design framework to optimize and implement the FIA monitors at any process node. Our design is fabricated and verified in 65 nm CMOS technology with a small footprint of 1500 um2. It can lock to clock frequencies from 2 MHz to 1.26 GHz while detecting all 12 types of possible clock glitches, as well as timing FIA injections via the supply voltage, electromagnetic signals, and chip temperature.

eess.SY↗