Jiangchang Xu

Self-supervised learning technology has been applied to calculate depth and ego-motion from monocular videos, achieving remarkable performance in various real-world scenarios. Unfortunately, challenges such as specular reflections and soft tissue deformations in endoscopic scenes greatly undermine the performance of these methods, inevitably compromising the accuracy of depth and ego-motion estimation. To address these two problems, we introduce a novel strategy based on image distance transform for robust self-supervised learning for monocular depth estimation, effectively handling specular reflections in endoscopic scenes. Furthermore, we propose a soft tissue deformation constraint based on biomechanical principles, which mitigates the adverse effects of deformed region pixels, ultimately enhancing the model’s depth estimation precision. Additionally, our method employs a lightweight architecture ensuring a reduced number of model parameters and faster inference time. Extensive experiments are conducted on both public datasets (SCARED, SERV-CT) and our own datasets to validate the effectiveness of our method. Compared with other SOTA methods, our approach demonstrates comparable accuracy and robustness while ensuring faster inference time. On the SCARED dataset, our approach attains an RMSE of 4.96 mm with only 2.25M model parameters for depth estimation. Especially, experiment results on SERV-CT dataset and our own datasets further demonstrate the model’s generalization ability and potential clinical value in computer-assisted surgical navigation.

Despite the advances in automated medical image segmentation, AI models still underperform in various clinical settings, challenging real-world integration. In this multicenter evaluation, we analyzed 20 state-of-the-art mandibular segmentation models across 19,218 segmentations of 1,000 clinically resampled CT/CBCT scans. We show that segmentation accuracy varies by up to 25% depending on socio-technical factors such as voxel size, bone orientation, and patient conditions such as osteosynthesis or pathology. Higher sharpness, isotropic smaller voxels, and neutral orientation significantly improved results, while metallic osteosynthesis and anatomical complexity led to significant degradation. Our findings challenge the common view of AI models as “plug-and-play” tools and suggest evidence-based optimization recommendations for both clinicians and developers. This will in turn boost the integration of AI segmentation tools in routine healthcare.

Positron emission tomography (PET) is one of the most advanced imaging diagnostic devices in the medical field, playing a crucial role in tumor diagnosis and treatment. However, patient motion during scanning can lead to motion artifacts, which affect diagnostic accuracy. This study aims to develop a head motion monitoring system to identify and select images with excessive motion and corresponding periods. The system, based on an RGB-D structured-light camera, implements facial feature point detection, 3D information acquisition, and head motion monitoring, along with a user interaction software. Through phantom experiments and volunteer experiments, the system’s performance was tested under various conditions, including stillness, pitch movement, yaw movement, and comprehensive movement. Experimental results show that the system’s translational error is less than 2.5 mm, rotational error is less than 2.0°, and it can output motion monitoring results within 10 s after the PET scanning, meeting clinical accuracy requirements and showing significant potential for clinical application.

Positron emission tomography (PET) is one of the most advanced imaging diagnostic devices in the medical field, playing a crucial role in tumor diagnosis and treatment. However, patient motion during scanning can lead to motion artifacts, which affect diagnostic accuracy. This study aims to develop a head motion monitoring system to identify and select images with excessive motion and corresponding periods. The system, based on an RGB-D structured light camera, implements facial feature point detection, 3D information acquisition, and head motion monitoring, along with a user interaction software. Through phantom experiments and volunteer experiments, the system’s performance was tested under various conditions, including stillness, pitch movement, yaw movement, and comprehensive movement. Experimental results show that the system’s translational error is less than 2.5 mm, rotational error is less than 2°, and it can output motion monitoring results within 10 seconds after the PET scan, meeting clinical accuracy requirements and showing significant potential for clinical application.

Maxillary Sinus Lifting is a crucial surgical procedure for addressing insufficient alveolar bone mass and severe resorption in dental implant therapy. To accurately analyze the geometry changes of the bone graft (BG) in the maxillary sinus (MS), it is essential to perform quantitative analysis. However, automated BG segmentation remains a major challenge due to the complex local appearance, including blurred boundaries, lesion interference, implant and artifact interference, and BG exceeding the MS. Currently, there are few tools available that can efficiently and accurately segment BG from cone beam computed tomography (CBCT) image. In this paper, we propose a distance-constrained attention network guided by prior anatomical knowledge for the automatic segmentation of BG. First, a guidance strategy of preoperative prior anatomical knowledge is added to a deep neural network (DNN), which improves its ability to identify the dividing line between the MS and BG. Next, a coordinate attention gate is proposed, which utilizes the synergy of channel and position attention to highlight salient features from the skip connections. Additionally, the geodesic distance constraint is introduced into the DNN to form multi-task predictions, which reduces the deviation of the segmentation result. In the test experiment, the proposed DNN achieved a Dice similarity coefficient of 85.48 $\pm$ 6.38%, an average surface distance error is 0.57 $\pm$ 0.34mm, and a 95% Hausdorff distance of 2.64 $\pm$ 2.09mm, which is superior to the comparison networks. It markedly improves the segmentation accuracy and efficiency of BG and has potential applications in analyzing its volume change and absorption rate in the future.

Purpose Orbital wall segmentation is critical for orbital measurement and reconstruction. However, the orbital floor and medial wall are made up of thin walls (TW) with low gradient values, making it difficult to segment the blurred areas of the CT images. Clinically, doctors have to manually repair the missing parts of TW, which is time-consuming and laborious. Methods To address these issues, this paper proposes an automatic orbital wall segmentation method based on TW region supervision using a multi-scale feature search network. First of all, in the encoding branch, the densely connected atrous spatial pyramid pooling based on the residual connection is adopted to achieve a multi-scale feature search. Then, for feature enhancement, multi-scale up-sampling and residual connection are applied to perform skip connection of features in multi-scale convolution. Finally, we explore a strategy for improving the loss function based on the TW region supervision, which effectively increases the TW region segmentation accuracy. Results The test results show that the proposed network performs well in terms of automatic segmentation. For the whole orbital wall region, the Dice coefficient (Dice) of segmentation accuracy reaches 96.086 ± 1.049%, the Intersection over Union (IOU) reaches 92.486 ± 1.924%, and the 95% Hausdorff distance (HD) reaches 0.509 ± 0.166 mm. For the TW region, the Dice reaches 91.470 ± 1.739%, the IOU reaches 84.327 ± 2.938%, and the 95% HD reaches 0.481 ± 0.082 mm. Compared with other segmentation networks, the proposed network improves the segmentation accuracy while filling the missing parts in the TW region. Conclusion In the proposed network, the average segmentation time of each orbital wall is only 4.05 s, obviously improving the segmentation efficiency of doctors. In the future, it may have a practical significance in clinical applications such as preoperative planning for orbital reconstruction, orbital modeling, orbital implant design, and so on.