SZShuai Zhang

Shuai Zhang

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Shuai Zhang has a total of 48 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • EndoLRMGS: Combining Large Reconstruction Modelling and Gaussian Splatting for Complete Endoscopic Scene Reconstruction

    by Xu Wang, Shuai Zhang, Baoru Huang, Jialang Xu, Danail Stoyanov, Evangelos B. Mazomenos

    Reconstructing dynamic surgical scenes from endoscopic videos remains a fundamental challenge in robot-assisted surgery. Existing methods primarily focus on deformable tissues, overlooking the presence of articulated instruments. To bridge this gap, we present EndoLRMGS, the first unified framework capable of reconstructing both deformable tissue and articulated instruments in a modular approach from monocular video and depth priors. We introduce Frequency-Modulated Gaussian Splatting (FMGS), for deformable tissue reconstruction, which modulates the spatial frequency of Gaussian primitives according to the Nyquist-Shannon sampling theorem, improving the visual fidelity while maintaining robust geometric accuracy. For instrument reconstruction, we leverage the Large Reconstruction Model (LRM) to generate high-quality, watertight 3D models from single images, and introduce a novel Orthographic and Perspective joint Projection Optimization (OPjPO) module to recover metric scale and spatial alignment. Extensive experiments on public datasets demonstrate the effectiveness of EndoLRMGS, achieving PSNR values from 28.4981 to 38.4179, with Chamfer distance ranging from 1.43 to 4.71 mm in tissue reconstruction. For instrument reconstruction, EndoLRMGS achieves PSNR values ranging from 19.7341 to 22.4393 on left views and 17.7442 to 20.8436 on right views. In terms of spatial alignment accuracy, it attains IoU values between 71.56% and 85.82%, with Chamfer distance ranging from 12.59 to 17.38 mm. These results highlight EndoLRMGS as a powerful and versatile solution for accurate, complete, and photorealistic 3D reconstruction of surgical scenes. Code is available at: EndoLRMGS.

  • 3D Acetabular Surface Reconstruction from 2D Pre-operative X-Ray Images Using SRVF Elastic Registration and Deformation Graph

    by Shuai Zhang, Jinliang Wang, Xu Wang, Sujith Konan, Danail Stoyanov, Evangelos B. Mazomenos

  • StereoMamba: Real-Time and Robust Intraoperative Stereo Disparity Estimation via Long-Range Spatial Dependencies

    by Xu Wang, Jialang Xu, Shuai Zhang, Baoru Huang, Danail Stoyanov, Evangelos B. Mazomenos

    Stereo disparity estimation is crucial for obtaining depth information in robot-assisted minimally invasive surgery (RAMIS). While current deep learning methods have made significant advancements, challenges remain in achieving an optimal balance between accuracy, robustness, and inference speed. To address these challenges, we propose the StereoMamba architecture, which is specifically designed for stereo disparity estimation in RAMIS. Our approach is based on a novel Feature Extraction Mamba (FE-Mamba) module, which enhances long-range spatial dependencies both within and across stereo images. To effectively integrate multi-scale features from FE-Mamba, we then introduce a novel Multidimensional Feature Fusion (MFF) module. Experiments against the state-of-the-art on the ex-vivo SCARED benchmark demonstrate that StereoMamba achieves superior performance on EPE of 2.64 px and depth MAE of 2.55 mm, the second-best performance on Bad2 of 41.49% and Bad3 of 26.99%, while maintaining an inference speed of 21.28 FPS for a pair of high-resolution images (1280 × 1024), striking the optimum balance between accuracy, robustness, and efficiency. Furthermore, by comparing synthesized right images, generated from warping left images using the generated disparity maps, with the actual right image, StereoMamba achieves the best average SSIM (0.8970) and PSNR (16.0761), exhibiting strong zero-shot generalization on the in-vivo RIS2017 and StereoMIS datasets.


  • Adjunct Tools for Colonoscopy Enhancement: A Comprehensive Review

    by Neri Niccolò Dei, Evangelos B. Mazomenos, Shuai Zhang, Sophia Bano, José M. M. Montiel, Danail Stoyanov, Gastone Ciuti

    Colonoscopy is considered the gold standard for detecting and diagnosing colorectal cancer (CRC), which is the second most common cause of cancer-related deaths worldwide. While colonoscopy is generally safe and effective at reducing CRC mortality, lesions can be missed during procedures, with adverse impacts on the patient. Latest innovations in hardware and software led to the development of adjunct tools for complementing standard colonoscopy to ensure optimal outcomes. Such tools aim to enhance the detection of lesions, standardize procedures, enhance safety, and minimize discomfort. Ultimately, they contribute to reducing the morbidity and mortality rates associated with CRC. This survey comprehensively explores both clinically tested and emerging advanced hardware and software adjunct tools, categorizing them based on their role in targeting three clinical challenges: mucosal visualization, lesion detection and classification, and navigation and procedure assessment. Moreover, this analysis allows exploring synergistic strategies for the future of the practice, with a focus on the promising role of AI-embedded robotic technologies.


  • Open Access

    3D Acetabular Surface Reconstruction from 2D Pre-operative X-ray Images using SRVF Elastic Registration and Deformation Graph

    by Shuai Zhang, Jinliang Wang, Sujith Konandetails, Xu Wang, Danail Stoyanov, Evangelos B. Mazomenos

    Accurate and reliable selection of the appropriate acetabular cup size is crucial for restoring joint biomechanics in total hip arthroplasty (THA). This paper proposes a novel framework that integrates square-root velocity function (SRVF)-based elastic shape registration technique with an embedded deformation (ED) graph approach to reconstruct the 3D articular surface of the acetabulum by fusing multiple views of 2D pre-operative pelvic X-ray images and a hemispherical surface model. The SRVF-based elastic registration establishes 2D-3D correspondences between the parametric hemispherical model and X-ray images, and the ED framework incorporates the SRVF-derived correspondences as constraints to optimize the 3D acetabular surface reconstruction using nonlinear least-squares optimization. Validations using both simulation and real patient datasets are performed to demonstrate the robustness and the potential clinical value of the proposed algorithm. The reconstruction result can assist surgeons in selecting the correct acetabular cup on the first attempt in primary THA, minimising the need for revision surgery.

  • Adaptive Dual-Axis Style-Based Recalibration Network With Class-Wise Statistics Loss for Imbalanced Medical Image Classification

    by Xiaoqing Zhang, Zunjie Xiao, Jingzhe Ma, Xiao Wu, Jilu Zhao, Shuai Zhang, Runzhi Li, Yi Pan, Jiang Liu

    Salient and small lesions (e.g., microaneurysms on fundus) both play significant roles in real-world disease diagnosis under medical image examinations. Although deep neural networks (DNNs) have achieved promising medical image classification performance, they often have limitations in capturing both salient and small lesion information, restricting performance improvement in imbalanced medical image classification. Recently, with the advent of DNN-based style transfer in medical image generation, the roles of clinical styles have attracted great interest, as they are crucial indicators of lesions. Motivated by this observation, we propose a novel Adaptive Dual-Axis Style-based Recalibration (ADSR) module, leveraging the potential of clinical styles to guide DNNs in effectively learning salient and small lesion information from a dual-axis perspective. ADSR first emphasizes salient lesion information via global style-based adaptation, then captures small lesion information with pixel-wise style-based fusion. We construct an ADSR-Net for imbalanced medical image classification by stacking multiple ADSR modules. Additionally, DNNs typically adopt cross-entropy loss for parameter optimization, which ignores the impacts of class-wise predicted probability distributions. To address this, we introduce a new Class-wise Statistics Loss (CWS) combined with CE to further boost imbalanced medical image classification results. Extensive experiments on five imbalanced medical image datasets demonstrate not only the superiority of ADSR-Net and CWS over state-of-the-art (SOTA) methods but also their improved confidence calibration results. For example, ADSR-Net with the proposed loss significantly outperforms CABNet50 by 21.39% and 27.82% in F1 and B-ACC while reducing 3.31% and 4.57% in ECE and BS on ISIC2018.


  • Direct Camera-Only Bundle Adjustment for 3-D Textured Colon Surface Reconstruction Based on Pre-Operative Model

    by Shuai Zhang, Liang Zhao, Shoudong Huang, Evangelos B. Mazomenos, Danail Stoyanov

    This paper addresses the problem of reconstructing textured colon surface maps using a sequence of monocular colonoscopic images together with a 3D colon mesh model that has been segmented in CT colonography. The problem is formulated as a direct bundle adjustment (BA) problem which simultaneously optimizes all camera poses and the intensity of vertices on the pre-operative mesh model. This optimization is achieved by maximizing photometric consistency among multiple views of 2D images and the pre-operative 3D mesh model.The key properties of our proposed direct BA formulation involve eliminating the need for reference image specification, data association (feature extraction and matching), and image depth information. Thus, the proposed method is particularly suitable for scenarios where distinct features and image depth are not available, such as 2D colonoscopic images.Furthermore, we have proven that solving the proposed direct BA using the Gauss-Newton (GN) algorithm has the merit of optimizing camera poses only, which is equivalent to optimizing camera poses and the intensities of 3D vertices on the mesh together. Thus, a direct camera-only BA algorithm is proposed and used for 3D textured colon reconstruction from textureless 2D colonoscopic images.Validations using simulation, phantom, and in-vivo datasets are performed to demonstrate the accuracy and feasibility of the proposed algorithm.


  • Gaussian Pancakes: Geometrically-Regularized 3D Gaussian Splatting for Realistic Endoscopic Reconstruction

    by Sierra Bonilla, Shuai Zhang, Dimitrios Psychogyios, Danail Stoyanov, Francisco Vasconcelos, Sophia Bano


  • SLAM-TKA: Simultaneously Localizing X-Ray Device and Mapping Pins in Conventional Total Knee Arthroplasty

    by Shuai Zhang, Liang Zhao, Shoudong Huang, Hua Wang, Qi Luo, Qi Hao, Danail Stoyanov

    This paper presents a novel simultaneous localisation and mapping (SLAM) technique, termed SLAM-TKA, for assisting total knee arthroplasty (TKA), a highly effective orthopaedic surgery that replaces arthritic or dysfunctional joint surfaces with knee prostheses. Our proposed SLAM algorithm uses information from a pre-operative tibia CT scan, intra-operative 2D X-ray images, and a trocar pin 3D mesh model to simultaneously localise the X-ray device and map the two trocar pins. Then, the estimated pins are used to evaluate the accuracy of the bone resection plane before the actual bone cutting, which plays a crucial role in precisely implanting the knee prostheses. To ensure high accuracy and robustness of the proposed SLAM algorithm, three energy terms are proposed and used together to align the edge observations of the tibia, fibula and pins on the intra-operative X-ray images and their corresponding pre-operative 3D mesh models in both 2D and 3D space. To enable the proposed iteration-based SLAM algorithm to be implemented in real-time such that the evaluation processing does not interrupt much on the workflow of TKA, the data association of edge correspondences matching and exhausted points-to-mesh distance calculation are pre-computed using the signed distance field method. Simulations are used to evaluate the accuracy and robustness of the proposed algorithm, and the experiments using in-vivo datasets from five patients demonstrate the high accuracy and efficiency in practice. The code and datasets are released at https://github.com/zsustc/SLAM-TKA.


  • Open Access

    Is it feasible to develop a supervised learning algorithm incorporating spinopelvic mobility to predict impingement in patients undergoing total hip arthroplasty?

    by Andreas Fontalis, Baixiang Zhao, Pierre Putzeys, Fabio Mancino, Shuai Zhang, Thomas Vanspauwen, Fabrice Glod, Ricci Plastow, Evangelos Mazomenos, Fares S. Haddad

    Aims: Precise implant positioning, tailored to individual spinopelvic biomechanics and phenotype, is paramount for stability in total hip arthroplasty (THA). Despite a few studies on instability prediction, there is a notable gap in research utilizing artificial intelligence (AI). The objective of our pilot study was to evaluate the feasibility of developing an AI algorithm tailored to individual spinopelvic mechanics and patient phenotype for predicting impingement.Methods: This international, multicentre prospective cohort study across two centres encompassed 157 adults undergoing primary robotic arm-assisted THA. Impingement during specific flexion and extension stances was identified using the virtual range of motion (ROM) tool of the robotic software. The primary AI model, the Light Gradient-Boosting Machine (LGBM), used tabular data to predict impingement presence, direction (flexion or extension), and type. A secondary model integrating tabular data with plain anteroposterior pelvis radiographs was evaluated to assess for any potential enhancement in prediction accuracy.Results: We identified nine predictors from an analysis of baseline spinopelvic characteristics and surgical planning parameters. Using fivefold cross-validation, the LGBM achieved 70.2% impingement prediction accuracy. With impingement data, the LGBM estimated direction with 85% accuracy, while the support vector machine (SVM) determined impingement type with 72.9% accuracy. After integrating imaging data with a multilayer perceptron (tabular) and a convolutional neural network (radiograph), the LGBM’s prediction was 68.1%. Both combined and LGBM-only had similar impingement direction prediction rates (around 84.5%).Conclusion: This study is a pioneering effort in leveraging AI for impingement prediction in THA, utilizing a comprehensive, real-world clinical dataset. Our machine-learning algorithm demonstrated promising accuracy in predicting impingement, its type, and direction. While the addition of imaging data to our deep-learning algorithm did not boost accuracy, the potential for refined annotations, such as landmark markings, offers avenues for future enhancement. Prior to clinical integration, external validation and larger-scale testing of this algorithm are essential.Cite this article: Bone Jt Open 2024;5(8):671–680.