QLQicheng Lao

Qicheng Lao

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Qicheng Lao has a total of 150 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • CAT-WSI: Context-Aware Trajectory Learning for Whole-Slide Breast Pathology Segmentation

    by Jiajun Qiu, Chaoran Zhang, Guangjing Yang, Chang Zheng, Xiaorong Zhong, Zhang Zhang, Ting Luo, Shaoting Zhang, Qicheng Lao, Jin Yin, et al.

  • Conditional Diffusion Prompting for Ambiguous Medical Image Segmentation

    by Hongkai Zhao, Jun Gao, Qingbo Kang, Qicheng Lao

  • Multi-VLM collaborated adaptive sampling for enhanced data pruning

    by Changfan Wang, Chaoran Zhang, Wei Xu, Huahui Yi, Kang Li, Zhicheng Zhao, Boyu Wang, Fei Su, Qicheng Lao

  • Open Access

    Development and External Validation of an Efficient Deep Learning Model for Lesion Segmentation and Subtyping of Hepatic Echinococcosis in Ultrasound Imaging

    by Zhu He, Jiajun Qiu, Chenlin Du, Jin Yin, Xuhui Zhang, Yelei Ren, Yifei Wang, Lamu Suolang, Chunyang Li, Zongjiu Zhang, et al.

    Background: Hepatic echinococcosis is a zoonotic parasitic disease common in remote, resource‐limited pastoral regions. It mainly appears as cystic echinococcosis (CE), due to Echinococcus granulosus, and alveolar echinococcosis (AE), caused by multilocular Echinococcus species. CE can lead to biliary colic or vascular compression, while AE often mimics liver cancer and exhibits high long‐term mortality. Although ultrasonography is the diagnostic method of choice, its accuracy is affected by imaging artifacts and operator variability. Conventional deep learning models are often computationally heavy and lack interpretability, limiting their use in underdeveloped areas. Integrating frequency‐ and spatial‐domain features in a lightweight framework offers a promising solution for efficient segmentation and classification of hepatic echinococcosis. The aim of this study is to develop efficient deep learning models for hepatic echinococcosis segmentation and classification, facilitating large‐scale screening with non‐invasive, portable ultrasound imaging. Methods: This study utilized a large ultrasound dataset to train and evaluate a deep learning model, consisting of 20,112 images from 4437 patients in Shiqu County, Sichuan Province, China, an endemic area for hepatic echinococcosis. To further assess the model's robustness, an external dataset comprising 3340 images from 1123 patients at West China Hospital of Sichuan University was used for additional testing. By enhancing the correlation of image features in both the frequency and spatial domains for hepatic echinococcosis ultrasound images, and incorporating segmentation features to assist the classification task, the developed model achieves high efficiency and lightness. Results: The proposed model achieved a Dice coefficient of 80.67% and 78.12% for segmentation, and classification accuracy of 90.10% and 80.96% on the internal and external test sets, respectively. Compared to the lightweight state‐of‐the‐art (SOTA) model, it improves inference speed by 43.48% and increases classification accuracy by 8.89% and 16.32% on internal and external test sets, respectively. Compared to the standard SOTA model, it is only 8% of its size but boosts inference speed by 821.37%, with classification accuracy improvements of 3.65% and 4.57% on internal and external test sets, respectively. Conclusions: The proposed model offers efficient and accurate hepatic echinococcosis diagnosis, with a lightweight design suitable for both resource‐limited and advanced clinical settings.

  • Open Access

    Learning Conceptual Text Prompts from Visual Regions of Interest for Medical Image Segmentation

    by Zhu He, Haoran Zhang, Wentao Zhang, Shen Zhao, Qiqi Liu, Xiaohu Wu, Qicheng Lao


  • Open Access

    Improving Medical Visual Reinforcement Fine-Tuning via Perception and Reasoning Augmentation

    by Guangjing Yang, ZhangYuan Yu, Ziyuan Qin, Xinyuan Song, Huahui Yi, Qingbo Kang, Jun Gao, Yiyue Li, Chenlin Du, Qicheng Lao

    While recent advances in Reinforcement Fine-Tuning (RFT) have shown that rule-based reward schemes can enable effective post-training for large language models, their extension to cross-modal, vision-centric domains remains largely underexplored. This limitation is especially pronounced in the medical imaging domain, where effective performance requires both robust visual perception and structured reasoning. In this work, we address this gap by proposing VRFT-Aug, a visual reinforcement fine-tuning framework tailored for the medical domain. VRFT-Aug introduces a series of training strategies designed to augment both perception and reasoning, including prior knowledge injection, perception-driven policy refinement, medically informed reward shaping, and behavioral imitation. Together, these methods aim to stabilize and improve the RFT process. Through extensive experiments across multiple medical datasets, we show that our approaches consistently outperform both standard supervised fine-tuning and RFT baselines. Moreover, we provide empirically grounded insights and practical training heuristics that can be generalized to other medical image tasks. We hope this work contributes actionable guidance and fresh inspiration for the ongoing effort to develop reliable, reasoning-capable models for high-stakes medical applications.

  • MLAE: Masked LoRA Experts for Parameter-Efficient Fine-Tuning

    by Junjie Wang, Guangjing Yang, Wentao Chen, Huahui Yi, Xiaohu Wu, Zhouchen Lin, Qicheng Lao

    In response to the challenges posed by the extensive parameter updates required for full fine-tuning of large-scale pre-trained models, parameter-efficient fine-tuning (PEFT) methods, exemplified by Low-Rank Adaptation (LoRA), have emerged. LoRA simplifies the fine-tuning process but may still struggle with a certain level of redundancy in low-rank matrices and limited effectiveness from merely increasing their rank. To address these issues, a natural idea is to enhance the independence and diversity of the learning process for the low-rank matrices. Therefore, we propose Masked LoRA Experts (MLAE), an innovative approach that applies the concept of masking to visual PEFT. Our method incorporates a cellular decomposition strategy that treats rank-1 components as experts defined under the chosen LoRA parameterization, thus enhancing diversity among update components. Additionally, we introduce a binary mask matrix that selectively activates these experts during training to promote more diverse and anisotropic learning, based on expert-level dropout strategies. Our investigations reveal that this selective activation not only enhances performance but also fosters a more diverse acquisition of knowledge with a marked decrease in parameter similarity among MLAE, significantly boosting the quality of the model. Remarkably, MLAE achieves new state-of-the-art (SOTA) performance with an average accuracy score of 78.8% on the VTAB-1k benchmark and 90.9% on the FGVC benchmark, surpassing the previous SOTA result by an average of 0.8% on both benchmarks. Moreover, MLAE shows strong generalization across diverse tasks, including LLM fine-tuning, semantic segmentation, and image/videotext understanding, underscoring its versatility and effectiveness in advancing PEFT.


  • Open Access

    Conda: Column-Normalized Adam for Training Large Language Models Faster

    by Junjie Wang, Pan Zhou, Yiming Dong, Huan Li, Jia Li, Xun Zhou, Qicheng Lao, Cong Fang, Zhouchen Lin

    Large language models (LLMs) have demonstrated impressive generalization and emergent capabilities, yet their pre-training remains computationally expensive and sensitive to optimization dynamics. While Adam-based optimizers offer fast convergence by adapting learning rates coordinate-wise, recent studies reveal that their updates often suffer from poor spectral conditioning and low-rank structures, hindering efficiency. Muon addresses this issue via global spectral normalization but lacks the per-coordinate adaptivity of Adam. In this work, we propose Column-Normalized Adam (Conda), a novel optimizer that bridges the strengths of both approaches. Conda projects updates into an orthogonal subspace and applies column-wise second moment normalization based on the projected gradients, thereby achieving both improved spectral conditioning and maintaining coordinate-wise adaptivity. This design alleviates the spectral pathologies of Adam while preserving its fast convergence behavior. Extensive experiments on the LLaMA and GPT-2 series show that Conda consistently outperforms AdamW, Muon, and other baselines in pre-training. Remarkably, on the LLaMA series, Conda achieves 2-2.5 the convergence speed of AdamW, measured in both training steps and training time. Further ablations demonstrate its robustness under diverse training setups. These results collectively highlight Conda as an effective and broadly applicable optimizer for large-scale LLM training. The code is released on https://github.com/jie040109/Conda

  • Endoscopic Artifact Inpainting for Improved Endoscopic Image Segmentation

    by Zhangyuan Yu, Chenlin Du, Hongrui Liang, Xiuqi Zheng, Zeyao Ma, Mingjun Wu, Mingwu Ao, Qicheng Lao