Doctoral Researcher in Ultrasound Medicine · Medical Imaging AI
浙江大学超声医学博士研究生 · 医学影像人工智能

I am a physician-researcher at Zhejiang University working at the intersection of clinical medicine, medical imaging, and artificial intelligence. My research begins with concrete clinical questions and emphasizes reproducible methods, explainable models, and evidence that can withstand scrutiny.

我是浙江大学超声医学专业博士研究生，关注临床医学、医学影像与人工智能的交叉研究。我的研究从具体临床问题出发，重视可复现的方法、可解释的模型，以及经得起审查的证据。

• Ultrasound AI / 超声人工智能 — thyroid, breast, contrast-enhanced ultrasound, and 3D reconstruction
• Medical imaging AI / 医学影像人工智能 — screening, prediction, and clinically grounded model validation
• Multimodal learning / 多模态学习 — integrating imaging, clinical variables, and structured medical knowledge
• Multi-center research / 多中心研究 — data heterogeneity, privacy-preserving learning, and reproducibility
• Explainable AI / 可解释人工智能 — understanding model evidence rather than reporting performance alone

Zhejiang University · 浙江大学
Doctoral researcher in Ultrasound Medicine, 2025–present
超声医学专业博士研究生，2025–至今

Qingdao University · 青岛大学
Clinical Medicine 5+3 Program; master's training in Pediatrics, 2017–2025
临床医学 5+3 一体化；硕士阶段主修儿科学，2017–2025

My earlier work focused on pediatric cardiovascular imaging AI, especially congenital heart disease screening and risk prediction from chest radiographs. My doctoral research is extending this experience toward ultrasound, multimodal modeling, and multi-center clinical data.

硕士阶段主要研究儿科心血管影像 AI，尤其是基于胸片的先天性心脏病筛查与风险预测；博士阶段进一步转向超声、多模态建模与多中心临床数据研究。

1. Zhixin Li, et al. CHD-CXR: a de-identified publicly available dataset of chest x-ray for congenital heart disease. Frontiers in Cardiovascular Medicine, 2024.

2. Zhixin Li, et al. Explanatory deep learning to predict elevated pulmonary artery pressure in children with ventricular septal defects using standard chest x-rays: a novel approach. Frontiers in Cardiovascular Medicine, 2024.

3. Zhixin Li, et al. The development and validation of an artificial intelligence-based screening method for atrial septal defect in children's chest x-rays. Frontiers in Pediatrics, 2023.

4. Gang Luo†, Zhixin Li†, et al. Residual networks models detection of atrial septal defect from chest radiographs. La Radiologia Medica, 2023. †Co-first authors.

5. Gang Luo†, Zhixin Li†, et al. An explainable deep learning model to predict partial anomalous pulmonary venous connection for patients with atrial septal defect. BMC Pediatrics, 2024. †Co-first authors.

• ST-Diff — research code exploring mask- and text-conditioned breast ultrasound image generation; currently an experimental prototype.
• ST-Diff — 基于病灶掩膜与临床文本条件的乳腺超声图像生成研究代码，目前为实验性原型。

• Veridrop Enhanced — self-hosted relay authenticity, protocol behavior, long-context, and reasoning evaluation.
• CandyBench CLI — a minimal terminal benchmark that preserves raw multi-model reasoning responses.

These tools reflect the same research principle I apply in medical AI: authenticity, availability, and reasoning ability are different measurements, and polished output is not a substitute for evidence.

这些工具延续了我在医学 AI 研究中的同一原则：真实性、可用性与推理能力是不同维度，流畅输出不能替代证据。

• National Scholarship, 2024 · 2024 年国家奖学金
• First-Class Graduate Scholarship, two years · 研究生一等奖学金（两年）
• Xinye Scholarship · 新业奖学金

Methods should be real, reproducible, and able to withstand peer scrutiny.
方法要真实、可复现、经得起同行审查。

• Start from a concrete clinical problem.

• Preserve raw evidence and state uncertainty honestly.

• Investigate the root cause of contradictions instead of hiding them.

• Avoid claims that exceed the available data.

• 从具体临床问题出发。

• 保留原始证据，如实说明不确定性。

• 面对矛盾追查根因，而不是掩盖问题。

• 不做超出数据证据边界的结论。

• Academic email / 学术邮箱: lizhixin@zju.edu.cn
• GitHub: @galaxy99881
• Location / 所在地: Hangzhou, Zhejiang, China · 中国浙江杭州

Python · PyTorch · Medical Imaging · Ultrasound · Explainable AI · Multimodal Learning · Reproducible Research