Abstract:In scoliosis surgery, the limited field of view of the C-arm X-ray machine restricts the surgeons' holistic analysis of spinal structures .This paper presents an end-to-end efficient and robust intraoperative X-ray image stitching method for scoliosis surgery,named SX-Stitch. The method is divided into two stages:segmentation and stitching. In the segmentation stage, We propose a medical image segmentation model named Vision Mamba of Spine-UNet (VMS-UNet), which utilizes the state space Mamba to capture long-distance contextual information while maintaining linear computational complexity, and incorporates the SimAM attention mechanism, significantly improving the segmentation <a class="link-external link-http" href="http://performance.In" rel="external noopener nofollow">this http URL</a> the stitching stage, we simplify the alignment process between images to the minimization of a registration energy function. The total energy function is then optimized to order unordered images, and a hybrid energy function is introduced to optimize the best seam, effectively eliminating parallax artifacts. On the clinical dataset, Sx-Stitch demonstrates superiority over SOTA schemes both qualitatively and quantitatively.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that during scoliosis surgery, due to the limited field of view of the C - arm X - ray machine, surgeons are unable to obtain a complete image of the spinal structure. To address this issue, the paper proposes an effective and robust intraoperative X - ray image stitching method named SX - Stitch. ### Problem Background During scoliosis surgery, the C - arm X - ray machine can only acquire partial image fragments of the spine because of its limited Field of View (FOV). These fragments need to be combined into a panoramic view through image stitching techniques so that surgeons can conduct a comprehensive analysis of the spinal structure. However, traditional image stitching methods face the following challenges when dealing with spinal X - ray images: 1. **Weak Features and Repeated Textures**: Spinal X - ray images have weak features and repeated textures, resulting in poor performance in terms of robustness for manually - designed features. 2. **Insufficient Data**: Deep learning methods usually require a large number of image pairs for training, but it is difficult to obtain sufficient data for effective model training in intraoperative truncated spinal X - ray images. 3. **Limitations of Existing Methods**: Traditional methods and existing deep learning methods perform well in processing natural - scene images, but they are not effective in the task of spinal X - ray image stitching. ### Solution To solve the above problems, the paper proposes the SX - Stitch framework, which consists of two main stages: 1. **Segmentation Stage**: - A medical image segmentation model named Vision Mamba of Spine - UNet (VMS - UNet) is proposed. This model uses the state - space Mamba to capture long - distance context information and combines the SimAM attention mechanism to significantly improve the segmentation performance. 2. **Stitching Stage**: - The alignment process between images is simplified and transformed into the problem of minimizing the registration energy function. The unordered images are sorted by optimizing the total energy function, and a hybrid energy function is introduced to optimize the best suture line, effectively eliminating parallax artifacts. ### Experimental Results The experimental results show that SX - Stitch outperforms the existing state - of - the - art (SOTA) schemes on clinical datasets and performs well in both qualitative and quantitative evaluations. Specifically: - **Segmentation Performance**: VMS - UNet performs well in segmenting the pedicle screw area with high accuracy. - **Stitching Quality**: SX - Stitch can accurately align image content, eliminate artifacts and distortion, and provide high - quality stitching results. - **Real - time Performance**: SX - Stitch can maintain efficient computational performance at different resolutions and overlap rates, with a short processing time and real - time capability. In conclusion, this paper aims to solve the problem of X - ray image stitching in scoliosis surgery through the SX - Stitch framework, providing surgeons with a more comprehensive and accurate view of the spinal structure, thereby increasing the success rate and safety of the surgery.

SX-Stitch: An Efficient VMS-UNet Based Framework for Intraoperative Scoliosis X-Ray Image Stitching

Three-Dimensional Stitching of Binocular Endoscopic Images Based on Feature Points

Automatic Spinal Ultrasound Image Segmentation and Deployment for Real-time Spine Volumetric Reconstruction

Self-Adaptive Ultrasound Scanning System for Imaging Human Spine

Studies on Image Stitching Algorithms in Machine Vision Inspection of Solar Panel

Improvement of automated image stitching system for DR X-ray images

Compliant Joint Based Robotic Ultrasound Scanning System for Imaging Human Spine

A novel multi-view X-ray digital imaging stitching algorithm

SynStitch: a Self-Supervised Learning Network for Ultrasound Image Stitching Using Synthetic Training Pairs and Indirect Supervision

Gastrointestinal image stitching based on improved unsupervised algorithm

Spine Vision X-Ray Image based GUI Planning of Pedicle Screws Using Enhanced YOLOv5 for Vertebrae Segmentation

Automatic intraoperative stitching of nonoverlapping cone-beam CT acquisitions

Adaptive loss engine for x-ray segmentation (ALEXS) for scoliosis intervention: assess digital segmentation and angle approximation

X23D—Intraoperative 3D Lumbar Spine Shape Reconstruction Based on Sparse Multi-View X-ray Data

Unsupervised Neural Network-Based Image Stitching Method for Bladder Endoscopy

The algorithm of stereo vision and shape from shading based on endoscope imaging

A Virtual Scanning Framework for Robotic Spinal Sonography with Automatic Real-time Recognition of Standard Views

SegStitch: Multidimensional Transformer for Robust and Efficient Medical Imaging Segmentation

Research on Panoramic Stitching Algorithm of Lateral Cranial Sequence Images in Dental Multifunctional Cone Beam Computed Tomography

2D Reconstruction of Small Intestine's Interior Wall

Multi-step 3D/2D Image Registration for Image-guided Spinal Surgery