Author name cluster

Yufan Chen

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

7 papers

1 author row

AAAI Conference 2026 Conference Paper

HybriDLA: Hybrid Generation for Document Layout Analysis

Yufan Chen
Omar Moured
Ruiping Liu
Junwei Zheng
Kunyu Peng
Jiaming Zhang
Rainer Stiefelhagen

Conventional document layout analysis (DLA) traditionally depends on empirical priors or a fixed set of learnable queries executed in a single forward pass. While sufficient for early-generation documents with a small, predetermined number of regions, this paradigm struggles with contemporary documents, which exhibit diverse element counts and increasingly complex layouts. To address challenges posed by modern documents, we present HybriDLA, a novel generative framework that unifies diffusion and autoregressive decoding within a single layer. The diffusion component iteratively refines bounding-box hypotheses, whereas the autoregressive component injects semantic and contextual awareness, enabling precise region prediction even in highly varied layouts. To further enhance detection quality, we design a multi-scale feature-fusion encoder that captures both fine-grained and high-level visual cues. This architecture elevates performance to 83.5% mean Average Precision (mAP). Extensive experiments on the DocLayNet and M6Doc benchmarks demonstrate that HybriDLA sets a state-of-the-art performance, outperforming previous approaches.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

HopaDIFF: Holistic-Partial Aware Fourier Conditioned Diffusion for Referring Human Action Segmentation in Multi-Person Scenarios

Kunyu Peng
Junchao Huang
Xiangsheng Huang
Di Wen
Junwei Zheng
Yufan Chen
Kailun Yang
Jiamin Wu

Action segmentation is a core challenge in high-level video understanding, aiming to partition untrimmed videos into segments and assign each a label from a predefined action set. Existing methods primarily address single-person activities with fixed action sequences, overlooking multi-person scenarios. In this work, we pioneer textual reference-guided human action segmentation in multi-person settings, where a textual description specifies the target person for segmentation. We introduce the first dataset for Referring Human Action Segmentation, i. e. , RHAS133, built from 133 movies and annotated with 137 fine-grained actions with 33h video data, together with textual descriptions for this new task. Benchmarking existing action segmentation methods on RHAS133 using VLM-based feature extractors reveals limited performance and poor aggregation of visual cues for the target person. To address this, we propose a holistic-partial aware Fourier-conditioned diffusion framework, i. e. , HopaDIFF, leveraging a novel cross-input gate attentional xLSTM to enhance holistic-partial long-range reasoning and a novel Fourier condition to introduce more fine-grained control to improve the action segmentation generation. HopaDIFF achieves state-of-the-art results on RHAS133 in diverse evaluation settings. The dataset and code are available at https: //github. com/KPeng9510/HopaDIFF.

PDF Details

NeurIPS Conference 2025 Conference Paper

Situat3DChange: Situated 3D Change Understanding Dataset for Multimodal Large Language Model

Ruiping Liu
Junwei Zheng
Yufan Chen
Zirui Wang
Kunyu Peng
Kailun Yang
Jiaming Zhang
Marc Pollefeys

Physical environments and circumstances are fundamentally dynamic, yet current 3D datasets and evaluation benchmarks tend to concentrate on either dynamic scenarios or dynamic situations in isolation, resulting in incomplete comprehension. To overcome these constraints, we introduce Situat3DChange, an extensive dataset supporting three situation-aware change understanding tasks following the perception-action model: 121K question-answer pairs, 36K change descriptions for perception tasks, and 17K rearrangement instructions for the action task. To construct this large-scale dataset, Situat3DChange leverages 11K human observations of environmental changes to establish shared mental models and shared situational awareness for human-AI collaboration. These observations, enriched with egocentric and allocentric perspectives as well as categorical and coordinate spatial relations, are integrated using an LLM to support understanding of situated changes. To address the challenge of comparing pairs of point clouds from the same scene with minor changes, we propose SCReasoner, an efficient 3D MLLM approach that enables effective point cloud comparison with minimal parameter overhead and no additional tokens required for the language decoder. Comprehensive evaluation on Situat3DChange tasks highlights both the progress and limitations of MLLMs in dynamic scene and situation understanding. Additional experiments on data scaling and cross-domain transfer demonstrate the task-agnostic effectiveness of using Situat3DChange as a training dataset for MLLMs. The established dataset and source code are publicly available at: https: //github. com/RuipingL/Situat3DChange.

PDF Details

NeurIPS Conference 2024 Conference Paper

Advancing Open-Set Domain Generalization Using Evidential Bi-Level Hardest Domain Scheduler

Kunyu Peng
Di Wen
Kailun Yang
Ao Luo
Yufan Chen
Jia Fu
M. Saquib Sarfraz
Alina Roitberg

In Open-Set Domain Generalization (OSDG), the model is exposed to both new variations of data appearance (domains) and open-set conditions, where both known and novel categories are present at test time. The challenges of this task arise from the dual need to generalize across diverse domains and accurately quantify category novelty, which is critical for applications in dynamic environments. Recently, meta-learning techniques have demonstrated superior results in OSDG, effectively orchestrating the meta-train and -test tasks by employing varied random categories and predefined domain partition strategies. These approaches prioritize a well-designed training schedule over traditional methods that focus primarily on data augmentation and the enhancement of discriminative feature learning. The prevailing meta-learning models in OSDG typically utilize a predefined sequential domain scheduler to structure data partitions. However, a crucial aspect that remains inadequately explored is the influence brought by strategies of domain schedulers during training. In this paper, we observe that an adaptive domain scheduler benefits more in OSDG compared with prefixed sequential and random domain schedulers. We propose the Evidential Bi-Level Hardest Domain Scheduler (EBiL-HaDS) to achieve an adaptive domain scheduler. This method strategically sequences domains by assessing their reliabilities in utilizing a follower network, trained with confidence scores learned in an evidential manner, regularized by max rebiasing discrepancy, and optimized in a bilevel manner. We verify our approach on three OSDG benchmarks, i. e. , PACS, DigitsDG, and OfficeHome. The results show that our method substantially improves OSDG performance and achieves more discriminative embeddings for both the seen and unseen categories, underscoring the advantage of a judicious domain scheduler for the generalizability to unseen domains and unseen categories. The source code is publicly available at https: //github. com/KPeng9510/EBiL-HaDS.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

FedPFT: Federated Proxy Fine-Tuning of Foundation Models

Zhaopeng Peng
Xiaoliang Fan
Yufan Chen
Zheng Wang
Shirui Pan
Chenglu Wen
Ruisheng Zhang
Cheng Wang

Adapting Foundation Models (FMs) for down- stream tasks through Federated Learning (FL) emerges a promising strategy for protecting data privacy and valuable FMs. Existing methods fine- tune FM by allocating sub-FM to clients in FL, however, leading to suboptimal performance due to insufficient tuning and inevitable error accumula- tions of gradients. In this paper, we propose Feder- ated Proxy Fine-Tuning (FedPFT), a novel method enhancing FMs adaptation in downstream tasks through FL by two key modules. First, the sub-FM construction module employs a layer-wise com- pression approach, facilitating comprehensive FM fine-tuning across all layers by emphasizing those crucial neurons. Second, the sub-FM alignment module conducts a two-step distillations—layer- level and neuron-level—before and during FL fine- tuning respectively, to reduce error of gradient by accurately aligning sub-FM with FM under theo- retical guarantees. Experimental results on seven commonly used datasets (i. e. , four text and three vi- sion) demonstrate the superiority of FedPFT. Our code is available at https: //github. com/pzp-dzd/FedPFT.

PDF Details DOI

YNIMG Journal 2022 Journal Article

Neurometabolic timecourse of healthy aging

Tao Gong
Steve C.N. Hui
Helge J. Zöllner
Mark Britton
Yulu Song
Yufan Chen
Aaron T. Gudmundson
Kathleen E. Hupfeld

PURPOSE: The neurometabolic timecourse of healthy aging is not well-established, in part due to diversity of quantification methodology. In this study, a large structured cross-sectional cohort of male and female subjects throughout adulthood was recruited to investigate neurometabolic changes as a function of age, using consensus-recommended magnetic resonance spectroscopy quantification methods. METHODS: 102 healthy volunteers, with approximately equal numbers of male and female participants in each decade of age from the 20s, 30s, 40s, 50s, and 60s, were recruited with IRB approval. MR spectroscopic data were acquired on a 3T MRI scanner. Metabolite spectra were acquired using PRESS localization (TE=30 ms; 96 transients) in the centrum semiovale (CSO) and posterior cingulate cortex (PCC). Water-suppressed spectra were modeled using the Osprey algorithm, employing a basis set of 18 simulated metabolite basis functions and a cohort-mean measured macromolecular spectrum. Pearson correlations were conducted to assess relationships between metabolite concentrations and age for each voxel; Spearman correlations were conducted where metabolite distributions were non-normal. Paired t-tests were run to determine whether metabolite concentrations differed between the PCC and CSO. Finally, robust linear regressions were conducted to assess both age and sex as predictors of metabolite concentrations in the PCC and CSO and separately, to assess age, signal-noise ratio, and full width half maximum (FWHM) linewidth as predictors of metabolite concentrations. RESULTS: Data from four voxels were excluded (2 ethanol; 2 unacceptably large lipid signal). Statistically-significant age*metabolite Pearson correlations were observed for tCho (r(98)=0.33, p 0.20). Age associations for tCho, tCr, mI and sI in the CSO and for NAAG, tCho, and tCr in the PCC remained when controlling for sex in robust regressions. CSO NAAG and Asp, as well as PCC tNAA, sI, and Lac were higher in women; PCC Gln was higher in men. When including an age*sex interaction term in robust regression models, a significant age*sex interaction was seen for tCho (F(1,96)=11.53, p=0.001) and GSH (F(1,96)=7.15, p=0.009) in the CSO and tCho (F(1,96)=9.17, p=0.003), tCr (F(1,96)=9.59, p=0.003), mI (F(1,96)=6.48, p=0.012), and Lac (F(1,78)=6.50, p=0.016) in the PCC. In all significant interactions, metabolite levels increased with age in females, but not males. There was a significant positive correlation between linewidth and age. Age relationships with tCho, tCr, and mI in the CSO and tCho, tCr, mI, and sI in the PCC were significant after controlling for linewidth and FWHM in robust regressions. CONCLUSION: The primary (correlation) results indicated age relationships for tCho, tCr, mI, and sI in the CSO and for NAAG, tCho, tCr, and Gln in the PCC, while no age correlations were found for tNAA, NAA, Glx, Glu, GSH, PE, Lac, or Asp in either region. Our results provide a normative foundation for future work investigating the neurometabolic time course of healthy aging using MRS.

Details DOI

YNICL Journal 2021 Journal Article

Focal corticarl dysplasia in epilepsy is associated with GABA increase

Tao Gong
Yubo Liu
Yufan Chen
Liangjie Lin
Youting Lin
Guangbin Wang

PURPOSE: Focal cortical dysplasia (FCD) is a major cause of drug-resistant epilepsy; however the underlying epileptogenic mechanisms of FCD metabolism in epilepsy patients remain unclear. The aim of this study is to detect alterations of γ-aminobutyric acid (GABA), glutathione (GSH), and the composite of glutamate and glutamine (Glx) in MRI-typical and neuropathologically confirmed FCD-associated epilepsy using Hadamard Encoding and Reconstruction of Mega-Edited Spectroscopy (HERMES). MATERIALS AND METHODS: Fourteen epileptic patients suspected to be caused by FCD and 14 healthy controls were enrolled prospectively in this study; all subjects underwent a 3 T MRI scan, including 3D T1 weighted imaging and HERMES. The GABA signal detected by HERMES also contains signals from macromolecules and homocarnosine, so it is referred as GABA+. Signals of GABA+, GSH and Glx detected by HERMES from tumor foci, contralateral cerebral regions, and healthy controls were quantified using Gannet. Fitting errors and signal to noise ratios (SNRs) of GABA + signals were also recorded. Differences of GABA+, GSH, Glx, fitting error and SNR of GABA + among three groups were analyzed using linear mixed effects models. RESULTS: Twelve FCD-associated epilepsy patients (7 females, aged 21.9 ± 9.3 years) and 12 matched healthy controls (7 females, aged 22.8 ± 9.8 years) were finally enrolled in this study. ANOVA results indicated that GABA levels were significantly increased in FCD foci compared with contralateral regions (p = 0.008) and with healthy controls (p = 0.003), while no difference was found in GSH and Glx levels. No difference of fitting errors or SNR of GABA + was found among FCD foci, contralateral regions and healthy controls. CONCLUSIONS: Increased GABA levels were found in FCD foci that indicated GABA may play a central role in the pathophysiology of FCD patients with epilepsy.

Details DOI