Weijie Ma

Weijie Ma

Auto-generated profile, based on publicly available information

Weijie Ma

Fudan University, Shanghai, China and Shanghai Innovation Institute, Shanghai, China

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
26 November 2025 in Nature
Verified
Journal Article
Open access
PDF file
WMWeijie Ma
Weijie Ma
ZZZhangrui Zhao
Zhangrui Zhao
CZCanyu Zhao
Canyu Zhao
WOWanli Ouyang
Wanli Ouyang
HZHan-Sen Zhong
Han-Sen Zhong
Glasses-free three-dimensional (3D) displays provide users with an immersive visual experience without the need of any wearable devices1,2. To achieve high-quality 3D imaging, a display should have both large linear dimensions and a wide viewing angle. However, the trade-off between spatial extent and bandwidth of optical systems, the space–bandwidth product, conventionally constrains the simultaneous maximization of the two. The two most common approaches to 3D displays are holographic3,4 and automultiscopic1,5,6, which, respectively, sacrifice either scale or viewing angle. Recently, some implementations enhanced by artificial intelligence have shown directions to mitigate these constraints, but they still operate within a set space–bandwidth product7,8. As a result, it remains challenging to fabricate large-scale wide-angle 3D displays9. Here we report the realization of a large-scale full-parallax 3D display with seamless viewing beyond 100°, maintained at over 50 Hz and 1,920 × 1,080 resolution on a low-cost light-field delivery setup. This device, called EyeReal, is realized by accurately modelling binocular view and combining it with a deep-learning real-time optimization, enabling the generation of optimal light-field outputs for each of the eyes. Our device could potentially enable applications in educational tools, 3D design and virtual reality10,11.

9 July 2026 in Research Square
Verified
Preprint
Open access
PDF file
WMWeijie Ma
Weijie Ma
WLWenjie Li
Wenjie Li
YZYujie Zhang
Yujie Zhang
FZFanrui Zhang
Fanrui Zhang
HSHaoran Sun
Haoran Sun
LWLei Wang
Lei Wang
Musculoskeletal diseases are among the leading causes of disability worldwide and generate the greatest global need for rehabilitation. Because recovery, remodelling and degeneration of bones, joints and related tissues often unfold over months to years, musculoskeletal care is fundamentally a longitudinal management problem rather than a single diagnostic or therapeutic decision. Effective care requires clinicians to repeatedly integrate evolving patient evidence, external medical knowledge and stage-specific functional goals. In routine practice, however, the evidence required for these decisions is often distributed across visits, departments and hospital information systems, breaking continuous care pathways into isolated episodes and limiting individualized, evidence-based management. Here we report OrthoPilot, a clinical artificial intelligence (AI) system powered by a large language model (LLM) that integrates hospital data streams with authoritative external knowledge for continuous musculoskeletal management. OrthoPilot autonomously retrieves real-time imaging, laboratory, pathology and order data, and converts evolving patient states into evidence-based decisions from admission diagnosis to rehabilitation planning. To benchmark full-pathway clinical management, we established a specialist-validated benchmark from real-world electronic health records (EHRs) spanning 1,000 disease codes. In a reader study, OrthoPilot was compared with 81 orthopaedic physicians across the complete care pathway, surpassing experts with 25 years of experience in diagnostic reasoning, clinical decision-making and management planning. Its advantage generalised across 60 external clinical centres, where it outperformed all evaluated intelligent systems. In a prospective physician decision-making study of 1,870 complex cases, OrthoPilot increased full-chain management success by 10.6%. In an 8-month randomised deployment involving 8,240 inpatients, OrthoPilot was embedded in routine hospital care, increasing cumulative cases per bed by 9.7% and raising patient-reported access to health information. Together, these results establish a new paradigm for clinical AI, moving from models that predict isolated events to systems that execute longitudinal management across complete musculoskeletal care pathways.


13 March 2025 in Research Square
Verified
Preprint
Open access
PDF file
HZHan-Sen Zhong
Han-Sen Zhong
WMWeijie Ma
Weijie Ma
ZZZhang-Rui Zhao
Zhang-Rui Zhao
WOWanli Ouyang
Wanli Ouyang
The pursuit of glasses-free three-dimensional (3D) displays represents an ultimate vision in display technology that could revolutionize human interaction with digital content. While near-eye solutions have shown promise at centimeter scales, scaling up to natural viewing circumstances (like desktop-monitor sizes) remains physically impossible due to the fundamental limitations of confined space-bandwidth product (SBP) in modern optical systems. Existing automultiscopic approaches attempt to address this challenge by trading viewing angles for larger imaging size, but under the limited SBP must compromise between broadish yet discrete effectual viewpoints, or continuous but restricted viewing zone. The limited SBP has made the display with both large scale and wide angle unattainable for over 50 years. Here we present EyeReal, the first artificial-intelligence-based method that maximally harness the constrained SBP to achieve true glasses-free 3D display over seamless ultrawide natural viewing ranges. The proposed neural network approach physically incorporates ocular geometric encoding and structured optical loss functions, enabling accurate but rapid computation of any-view optimal light field with physical-depth interpretability. Through extensive experimental validations, EyeReal demonstrates unprecedented 3D display with seamless viewing angles exceeding 100 degrees with complete parallax types (horizontal, vertical, radial, continuous, and focal) and no vergence-accommodation conflict, a combination previously thought impossible. This is maintained at high frame rates (over 50 hertz) with 1920×1080 resolutions on consumer hardware. This breakthrough establishes a potential foundation for next-generation consumer-grade 3D displays, opening new horizons across education, design, virtual interaction.