Author name cluster

Qiankun Li

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.

6 papers

1 author row

JBHI Journal 2025 Journal Article

A Reliable Multi-Stage System for Tooth Instance Segmentation and Numbering in Panoramic Radiographs

Yani Zhang
Yuzhou Yu
Qiankun Li
Huabao Chen
Yuci Liang
Junxin Chen
Jiayue Yin

Accurate tooth instance segmentation and identification in panoramic radiographs is a fundamental prerequisite for automated dental diagnosis and treatment planning. Existing methods often rely on static segmentation pipelines that overlook anatomical symmetry and fail to address structural anomalies such as supernumerary or missing teeth. To this end, we propose a reliable multi-stage framework for clinically applicable tooth instance segmentation and FDI-compliant numbering. The system comprises three stages: a supernumerary tooth classifier to guide adaptive routing, a core instance segmentor for tooth-level delineation and numbering, and a missing tooth detector to ensure index completeness via post-hoc correction. Central to our design is the Symmetry-Aware Dual-branch Pyramid Network (SADP-Net), which explicitly models bilateral structures and scale variations through a Symmetry-Aware Module (SymAM) and a Dual-Branch Pyramid (DBP) architecture. Extensive experiments on three datasets, including Tufts Dental, O 2 PR, and a multi-pathology cohort, demonstrate superior performance compared to state-of-the-art baselines. Ablation studies further validate the contributions of each proposed component in enhancing boundary localization, numbering consistency, and robustness to anatomical variability. Our framework provides a scalable and interpretable solution for real-world dental imaging systems. Codes link: https://github.com/qklee-lz/SADP-Net.

Details DOI

EAAI Journal 2025 Journal Article

Decoding text from electroencephalography signals: A novel Hierarchical Gated Recurrent Unit with Masked Residual Attention Mechanism

Qiupu Chen
Yimou Wang
Fenmei Wang
Duolin Sun
Qiankun Li

Progress in both neuroscience and natural language processing has opened doors for investigating brain to text techniques to reconstruct what individuals see, perceive, or focus on from human brain activity patterns. Non-invasive decoding, utilizing electroencephalography (EEG) signals, is preferred due to its comfort, cost-effectiveness, and portability. In brain-to-text applications, a pressing need has arisen to develop effective models that can accurately capture the intricate details of EEG signals, such as global and local contextual information and long-term dependencies. In response to this need, we propose the Hierarchical Gated Recurrent Unit with Masked Residual Attention Mechanism (HGRU-MRAM) model, which ingeniously combines the hierarchical structure and the masked residual attention mechanism to deliver a robust brain-to-text decoding system. Our experimental results on the ZuCo dataset demonstrate that this model significantly outperforms existing baselines, achieving state-of-the-art performance with Bilingual Evaluation Understudy Score (BLEU), Recall-Oriented Understudy for Gisting Evaluation (ROUGE), US National Institute of Standards and Technology Metric (NIST), Metric for Evaluation of Translation with Explicit Ordering (METEOR), Translation Edit Rate (TER), and BiLingual Evaluation Understudy with Representations from Transformers (BLEURT) scores of 48. 29, 34. 84, 4. 07, 34. 57, 21. 98, and 40. 45, respectively. The code is available at https: //github. com/qpuchen/EEG-To-Sentence.

Details DOI

NeurIPS Conference 2025 Conference Paper

From Pixels to Views: Learning Angular-Aware and Physics-Consistent Representations for Light Field Microscopy

Feng He
Guodong Tan
Qiankun Li
Jun Yu
Quan Wen

Light field microscopy (LFM) has become an emerging tool in neuroscience for large-scale neural imaging in vivo, with XLFM (eXtended Light Field Microscopy) notable for its single-exposure volumetric imaging, broad field of view, and high temporal resolution. However, learning-based 3D reconstruction in XLFM remains underdeveloped due to two core challenges: the absence of standardized datasets and the lack of methods that can efficiently model its angular–spatial structure while remaining physically grounded. We address these challenges by introducing three key contributions. First, we construct the XLFM-Zebrafish benchmark, a large-scale dataset and evaluation suite for XLFM reconstruction. Second, we propose Masked View Modeling for Light Fields (MVM-LF), a self-supervised task that learns angular priors by predicting occluded views, improving data efficiency. Third, we formulate the Optical Rendering Consistency Loss (ORC Loss), a differentiable rendering constraint that enforces alignment between predicted volumes and their PSF-based forward projections. On the XLFM-Zebrafish benchmark, our method improves PSNR by 7. 7\% over state-of-the-art baselines. Code and datasets are publicly available at: https: //github. com/hefengcs/XLFM-Former.

PDF Details

NeurIPS Conference 2025 Conference Paper

Unleashing Foundation Vision Models: Adaptive Transfer for Diverse Data-Limited Scientific Domains

Qiankun Li
Feng He
Huabao Chen
Xin Ning
Kun Wang
Zengfu Wang

In the big data era, the computer vision field benefits from large-scale datasets such as LAION-2B, LAION-400M, and ImageNet-21K, Kinetics, on which popular models like the ViT and ConvNeXt series have been pre-trained, acquiring substantial knowledge. However, numerous downstream tasks in specialized and data-limited scientific domains continue to pose significant challenges. In this paper, we propose a novel Cluster Attention Adapter (CLAdapter), which refines and adapts the rich representations learned from large-scale data to various data-limited downstream tasks. Specifically, CLAdapter introduces attention mechanisms and cluster centers to personalize the enhancement of transformed features through distribution correlation and transformation matrices. This enables models fine-tuned with CLAdapter to learn distinct representations tailored to different feature sets, facilitating the models' adaptation from rich pre-trained features to various downstream scenarios effectively. In addition, CLAdapter's unified interface design allows for seamless integration with multiple model architectures, including CNNs and Transformers, in both 2D and 3D contexts. Through extensive experiments on 10 datasets spanning domains such as generic, multimedia, biological, medical, industrial, agricultural, environmental, geographical, materials science, out-of-distribution (OOD), and 3D analysis, CLAdapter achieves state-of-the-art performance across diverse data-limited scientific domains, demonstrating its effectiveness in unleashing the potential of foundation vision models via adaptive transfer. Code is available at https: //github. com/qklee-lz/CLAdapter.

PDF Details

JBHI Journal 2024 Journal Article

Embracing Large Natural Data: Enhancing Medical Image Analysis via Cross-Domain Fine-Tuning

Qiankun Li
Xiaolong Huang
Bo Fang
Huabao Chen
Siyuan Ding
Xu Liu

With the rapid advancements of Big Data and computer vision, many large-scale natural visual datasets are proposed, such as ImageNet-21K, LAION-400M, and LAION-2B. These large-scale datasets significantly improve the robustness and accuracy of models in the natural vision domain. However, the field of medical images continues to face limitations due to relatively small-scale datasets. In this article, we propose a novel method to enhance medical image analysis across domains by leveraging pre-trained models on large natural datasets. Specifically, a Cross-Domain Transfer Module (CDTM) is proposed to transfer natural vision domain features to the medical image domain, facilitating efficient fine-tuning of models pre-trained on large datasets. In addition, we design a Staged Fine-Tuning (SFT) strategy in conjunction with CDTM to further improve the model performance. Experimental results demonstrate that our method achieves state-of-the-art performance on multiple medical image datasets through efficient fine-tuning of models pre-trained on large natural datasets.

Details DOI

AAAI Conference 2024 Conference Paper

One Step Learning, One Step Review

Xiaolong Huang
Qiankun Li
Xueran Li
Xuesong Gao

Visual fine-tuning has garnered significant attention with the rise of pre-trained vision models. The current prevailing method, full fine-tuning, suffers from the issue of knowledge forgetting as it focuses solely on fitting the downstream training set. In this paper, we propose a novel weight rollback-based fine-tuning method called OLOR (One step Learning, One step Review). OLOR combines fine-tuning with optimizers, incorporating a weight rollback term into the weight update term at each step. This ensures consistency in the weight range of upstream and downstream models, effectively mitigating knowledge forgetting and enhancing fine-tuning performance. In addition, a layer-wise penalty is presented to employ penalty decay and the diversified decay rate to adjust the weight rollback levels of layers for adapting varying downstream tasks. Through extensive experiments on various tasks such as image classification, object detection, semantic segmentation, and instance segmentation, we demonstrate the general applicability and state-of-the-art performance of our proposed OLOR. Code is available at https://github.com/rainbow-xiao/OLOR-AAAI-2024.

PDF Details DOI