Laparoscopic navigation systems increasingly leverage augmented reality (AR) to overlay preoperative anatomical data in real time, enhancing surgical precision. However, reliable navigation requires accurate camera pose estimation and high-fidelity 3D organ reconstruction. To address these challenges, we propose LA-SLAM, a real-time visual SLAM system specifically designed for laparoscopic surgery. The system incorporates three key innovations: 1) An optical-flow-based depth-pose joint estimation module. This module establishes accurate dense correspondences between images through the introduced global correlation softmax, providing sufficiently strong constraints for subsequent optimization to achieve high-precision pose and depth estimation with only a single update, significantly improving computational efficiency; 2) a hybrid loop closure strategy that integrates projection flow and dense optical flow for reliable detection, and introduces a Sim(3)-based optimization using 3D point cloud feature matching; and 3) a refined 3D reconstruction strategy combining depth consistency validation with statistical and gradient-based filtering to address laparoscopic challenges, including specular reflections, occlusions, and tissue deformation. Evaluations on standard SLAM datasets (TUM-RGBD, EuRoC) and laparoscopic datasets (SCARED, DePoLL, StereoMIS) demonstrate that LA-SLAM achieves tracking accuracy comparable to state-of-the-art SLAM systems while maintaining significant advantages in computational efficiency. It also produces geometrically faithful and visually robust reconstructions. These results highlight LA-SLAM's potential for integration into real-time surgical navigation systems.

Liver vessel segmentation has long been a challenging task in medical image segmentation because of the vessels' intricate branching patterns, varying diameters, low contrast against surrounding tissues, and the presence of very small regions that are hard to distinguish. To overcome these difficulties, we introduce multisampling feature fusion recover Unet (MFFR-Unet), an improved 3D-Unet tailored for small-vessel delineation. The core innovation is a multi-feature fusion module that amplifies vascular features through complementary downsampling strategies and refines them with channel-wise attention, thereby mitigating poor accuracy in tiny-vessel areas. This module is embedded in every skip connection of the 3D-Unet to act as a learnable feature selector. Preprocessing employs windowing and a slab-based cropping strategy, while training is driven by a composite Dice + BCE loss that counters severe class imbalance. A local confidence convolution repair (LCCR) module finally re-evaluates uncertain voxels using coarse probability maps and decoder features, yielding sharper boundaries. Trained and validated on the public 3Diradb liver-vessel set and the BraTS 2018 brain-tumor set, MFFR-Unet reaches a Dice score of 77.7% on 3Diradb an absolute 7% gain over the standard 3D-Unet and an average Dice of 82.7% on BraTS, surpassing most existing approaches. These results demonstrate that MFFR-Unet, through advanced feature extraction, effective preprocessing, and loss design, substantially enhances segmentation accuracy for challenging vascular structures in clinical imaging.

Laparoscopic navigation systems increasingly leverage augmented reality (AR) to overlay preoperative anatomical data in real time, enhancing surgical precision. However, reliable navigation requires accurate camera pose estimation and highfidelity 3D organ reconstruction. To address these challenges, we propose LA-SLAM, a real-time visual SLAM system specifically designed for laparoscopic surgery. The system incorporates three key innovations: (1) An optical-flow-based depth-pose joint estimation module. This module establishes accurate dense correspondences between images through the introduced global correlation softmax, providing sufficiently strong constraints for subsequent optimization to achieve high-precision pose and depth estimation with only a single update, significantly improving computational efficiency; (2) a hybrid loop closure strategy that integrates projection flow and dense optical flow for reliable detection, and introduces a Sim(3)-based optimization using 3D point cloud feature matching; and (3) a refined 3D reconstruction strategy combining depth consistency validation with statistical and gradient-based filtering to address laparoscopic challenges, including specular reflections, occlusions, and tissue deformation. Evaluations on standard SLAM datasets (TUM-RGBD, EuRoC) and laparoscopic datasets (SCARED, DePoLL, Stereo-MIS) demonstrate that LA-SLAM achieves tracking accuracy comparable to state-of-the-art SLAM systems while maintaining significant advantages in computational efficiency. It also produces geometrically faithful and visually robust reconstructions. These results highlight LA-SLAM’s potential for integration into real-time surgical navigation systems.

Laparoscopic liver surgery is a newly developed minimally invasive technique and represents an inevitable trend in the future development of surgical methods. By using augmented reality (AR) technology to overlay preoperative CT models with intraoperative laparoscopic videos, surgeons can accurately locate blood vessels and tumors, significantly enhancing the safety and precision of surgeries. Point cloud registration technology is key to achieving this effect. However, there are two major challenges in registering the CT model with the point cloud surface reconstructed from intraoperative laparoscopy. First, the surface features of the organ are not prominent. Second, due to the limited field of view of the laparoscope, the reconstructed surface typically represents only a very small portion of the entire organ. To address these issues, this paper proposes the keypoint correspondence registration network (KCR-Net). This network first uses the neighborhood feature fusion module (NFFM) to aggregate and interact features from different regions and structures within a pair of point clouds to obtain comprehensive feature representations. Then, through correspondence generation, it directly generates keypoints and their corresponding weights, with keypoints located in the common structures of the point clouds to be registered, and corresponding weights learned automatically by the network. This approach enables accurate point cloud registration even under conditions of extremely low overlap. Experiments conducted on the ModelNet40, 3Dircadb, DePoLL demonstrate that our method achieves excellent registration accuracy and is capable of meeting the requirements of real-world scenarios.

BACKGROUND AND OBJECTIVE:Virtual reality-based neurosurgical simulators show increasing potential for surgical training and preoperative planning. However, existing cutting models often lack a physics-based rupture mechanism and fail to preserve the biomechanical characteristics of soft tissue incisions, limiting their applicability to dura mater cutting simulations. Given the thin structure of the dura mater and its proximity to soft brain tissues, predicting rupture occurrences is essential. To address these challenges, this study presents a novel dura mater cutting model that enhances rupture prediction, incision realism, and simulation visualization. METHODS:A novel approach integrating Position-Based Dynamics with fracture mechanics to model dura mater cutting is introduced. The model employs the von Mises stress threshold criterion to estimate rupture initiation and location. Incision smoothing and mass redistribution techniques are employed to enhance incision geometry while preserving mass conservation. Additionally, adaptive constraints are used to simulate the post-cutting shrinkage effect of the dura mater. RESULTS:The presented model provides stable and realistic results during dura mater cutting simulation. The physics-based rupture results achieved are well aligned with those from ABAQUS finite element software, with a maximum discrepancy of 11.3%, while enabling real-time fracture prediction. Validation through stress distribution and cutting force analysis confirms the accuracy of the model, which in the meantime preserves computational efficiency and supports interactive visualization. Furthermore, the incision optimization and shrinkage simulation reproduce smooth incision and characteristic shrinkage behavior of the dura mater. CONCLUSIONS:The presented model offers a new approach to the simulation of dura mater cutting, integrating rupture prediction, incision optimization, and post-incision shrinkage. This is very important and meaningful in virtual neurosurgical simulation, enhancing surgical training and planning.

Deep ultraviolet (DUV) photodetectors in key fields increasingly demand higher capabilities for processing sensed information. In-sensor computing (ISC) has emerged as a revolutionary approach to enable perception beyond the sensing capability of traditional photodetectors. Ultrawide bandgap Ga2O3, with high DUV sensitivity and excellent stability, offers a subversive scheme for advanced DUV detection. Moreover, its conductivity plasticity, enabled by the persistent photoconductivity (PPC) in its photodetectors, lays the foundation for the application in DUV ISC systems. However, the PPC simultaneously damages the reproducibility for consecutive perception, while frequent computing increases heat production, particularly in integrated Ga2O3 photodetector chips. In this work, two-terminal Ga2O3/SiC photosynapses were tailored with reproducible plasticity and enhanced heat dissipation. A tailored bias Pulse Reset strategy is capable of erasing the PPC effect within 4.10 μs based on tunable tunneling effect at the interface, thus ensuring reproducibility in high-frequency temporal information processing. The introduced commercial SiC platform enhances heat dissipation. The high responsivity (15 A/W), in situ 4-bit encoding capability, and swift Reset speed rank this device among the top-tier DUV photodetectors and photosynapses. Furthermore, real-time perception and tracking of moving objects have been demonstrated based on this device in a residual neural network for ISC. This work offers the hetero-integrated Ga2O3/SiC photosynapses with enhanced reproducibility and heat dissipation, paving the way for efficient and scalable DUV ISC.

Through the integration of sensing and computing functions into a single photosynapse, the neuromorphic visual system mitigates the substantial data redundancy caused by frequent data conversion and transmission in Von Neumann architectures. However, most reported photosynapses can produce unidirectional light responses only without electric modulation and are limited to narrow spectral ranges, which limits their effectiveness in target recognition in complex real-world optical scenes. Here, we present a four-color reservoir computing (RC) system based on an opposite photogating (OPG)-engineered multispectral photosynapse. The OPG effect, characterized by light-modulated oppositely shifted threshold voltage (Vth), originates from different carrier dynamics in a Ga2O3/WSe2 heterojunction field-effect transistor. Specifically, hole trapping in Ga2O3 under deep ultraviolet (DUV) light induces negative Vth shifts (excitatory responses), while electron trapping in WSe2 under visible light causes positive Vth shifts (inhibitory responses). The nonlinear photoresponse and tunable short-term memory under external light stimuli make the photosynapse suitable for photoelectric reservoirs. The DUV-specific corona discharge, a critical challenge in high-voltage transmission systems, causes exacerbated equipment aging and significant energy losses. By integration of DUV-specific discharge signals and visible environmental information, the system achieves 88.3% accuracy in localizing the corona discharge among six high-risk components in high-voltage systems. Our multispectral RC system demonstrates a pathway toward precise intelligent image recognition in real-world multispectral scenarios.

Soft x-ray detectors play crucial roles in biology, chemistry, and lithography. Current soft x-ray detectors suffer from insufficient responsivity (R), excessively large cell area, and limited stability. Here, the β-Ga2O3 soft x-ray detector is constructed, and the effects of varying the sensitive layer thickness and voltage on the soft x-ray detection characteristics of the device are explored. Meanwhile, the mechanism of the multiplication ionization process from soft x-ray and the photoconductivity gain on the photoresponse performance of the device are analyzed. The device obtains the R up to 3.05 × 103 A/W under 300 eV soft x-ray irradiation at the synchrotron beamline, which is about 1.19 × 104 times higher than that of the conventional device. The Ga2O3 device also maintains stable operation under long-term irradiation and multicycle switching. These results indicate that Ga2O3 is an ideal candidate material for soft x-ray detection, which has great potential for applications such as imaging of biological cells.

Accurately registering point clouds is challenging due to three primary reasons: 1) it is difficult for point cloud feature descriptors to handle noise in complex scenes; 2) poorly descriptive features lead to incorrect sets of corresponding points; and 3) non-overlapping regions in the scene can adversely affect registration results. To address these issues, our approach consists of three key contributions. First, we propose a segmenting sphere region (SSR) feature descriptor that comprehensively preserves point cloud spatial coordinate information through the "sphere segmentation-furthest point preservation" operation, enabling robust registration in complex scenarios. Second, we design SSR-Net to improve the descriptiveness of SSR features, generating a soft matching matrix to estimate the correspondence between the improved features. Finally, we design an overlap region estimation module in SSR-Net, which employs attention to find the overlap region, thereby reducing the negative impact of non-overlapping regions in the soft matching matrix on registration results. We conducted comprehensive experiments on the B3R, ModelNet40, KITTI, unmanned aerial vehicle (UAV), and 3DMatch datasets, demonstrating the effectiveness of our proposed method.

Point cloud completion aims to predict the missing part for an incomplete 3D shape. Existing point cloud completion methods based on deep learning complete the point cloud by extracting global features from the incomplete point cloud. However, such methods cannot generate a uniformly distributed point cloud and the accurate structure details of the object. To solve the problem, a novel method for completing point clouds is proposed in this paper. Our approach is a two-step strategy. First, to predict the sparse point cloud with uniform density, the Sequence Generation Completion (SGC) method is proposed. By numbering the subspace obtained from the spatial subdivision, the point cloud is represented with a sequence of numbers and the point cloud completion problem is turned into a sequence generation problem. Second, to obtain the dense point cloud and generate the accurate structural details of point clouds, we propose a resolution scale network (RSN). This network takes local resolution as input and increases the weight of low-resolution regions by learning to preserve the comprehensive structural information of the sparse point cloud, which is crucial to generate dense point cloud. The comprehensive experiments on several public datasets demonstrate the effectiveness of our method. Source code and pretrained models will be available at github.com/Pikachu-NCU/Sequence-Generate-Completion-Method.

Ga2O3 has been considered as one of the most suitable materials for x-ray detection, but its x-ray detection performance is still at a low level due to the limitation of its quality and absorbance, especially for hard x-ray. In this work, the effects of growth temperature and miscut angle of the sapphire substrate on the crystal quality of Ga2O3 thin films were investigated based on the MOCVD technique. It was found that the crystal growth mode was transformed from island growth to step-flow growth using miscut sapphire substrates and increasing growth temperature, which was accompanied by the improvement of the crystal quality and the reduction of the density of trapped states. Ga2O3 films with optimal crystal quality were finally prepared on a 4 degrees miscut substrate at 900 degrees C. The x-ray detector based on this film shows good hard x-ray response with a sensitivity of 3.72 x 10(5)mu CGy(air)(-1)cm(-2). Furthermore, the impacts of Ga2O3 film crystal quality and trap density on the x-ray detector were investigated in depth, and the mechanism of the photoconductive gain of the Ga2O3 thin-film x-ray detector was analyzed.

Gallium oxide (Ga2O3), with an ultrawide bandgap corresponding to the deep ultraviolet (DUV) range, has attracted significant attention in optical filter-free photodetectors. In practical terms, DUV photodetectors employed in extreme conditions, for example, flame detection and space exploration, face the challenges of performance degradation caused by high/low-temperature transformation. Here, DUV photodetectors are tailored with high durability and stability by one-step-grown beta-Ga2O3 films via pulsed laser deposition. A high-oxygen-pressure scheme effectively addresses the issue of film-free deposition at specifically high temperatures, facilitating the formation of polycrystalline high-resistivity beta-Ga2O3 films. As a result, the devices exhibit outstanding performance, including a low dark current (4.4 pA @30 V), high photoresponsivity (147.36 A W-1), and fast response time (3.1/22.6 ms). Additionally, the photoresponse performance shows minimal degradation at high temperatures up to 300 degrees C and even improves at low temperatures down to -100 degrees C, ranking it among the most robust DUV photodetectors. The mechanism of photoresponse, involving the exciton formation, bandgap evolution, carrier-phonon scatter, etc., is also elucidated in a wide temperature range. This work provides an efficient solution for developing robust Ga2O3 DUV photodetectors with excellent performance for extreme-condition applications.

Brain tumor segmentation supplies a credible reference for clinical treatment and pathological research and facilitates practitioners to diagnose more accurately. However, since the randomness and complexity of tumor shape and location, automatic brain tumor segmentation remains an extremely challenging assignment. In this study, we build an end-to-end convolutional neural network with a U-shaped structure to implement the segmentation of three lesion regions. We propose a multi-scale context block and an attention guidance block to focus on the spatial information at different scales and the interdependence between feature channels to enhance network representations and boost the learning capability of the model. Specifically, the multi-scale context block draws rich feature information through 3D dilated convolution. The attention guidance block reduces the impact of learned redundant features and eliminates the interference of irrelevant regions in the overall global information. Our recommended approach is evaluated on the brain tumor segmentation 2020 validation data. The Dice scores of the enhancing tumor (ET), whole tumor (WT), and tumor core (TC) are 78.19%, 90.10%, and 83.98%, respectively. In addition, the practice is also carried out in 2019 online validation data, and the Dice scores of ET, WT, and TC are 77.31%, 89.64%, and 82.55%, respectively. Experimental results reveal that the recommended approach gains favorable performance in comparison with representative brain tumor segmentation approaches. Our present study would accurately and efficaciously segment the three brain lesion regions and has clinical practice value.

Abstract Deep ultraviolet (DUV) photodetectors play important roles in the modern semiconductor industry due to their diverse applications in critical fields. Wide bandgap semiconductor Ga2O3 is considered as one promising material for highly sensitive DUV photodetectors. However, the high responsivity of Ga2O3 DUV photodetectors always comes at the expense of its response speed. Material engineering for high-quality Ga2O3 materials can optimize the photoresponse performance but at the cost of much more complex process. Structure engineering can efficiently improve the performance of Ga2O3 photodetectors based on various physical mechanisms. Owing to the increased modulation probabilities, part schemes of structure engineering even alleviate the tough requirements on Ga2O3 material quality for high-performance DUV photodetectors. This article reviews the recent efforts in optimizing the performance of Ga2O3 photodetectors through structure engineering. Firstly, photodetectors based on Ga2O3 nanostructures and metasurface structures with nanometer size effect are discussed. In addition, junction structures of Ga2O3 photodetectors, which effectively promote carrier separation in the depletion region, are summarized based on a classification of Schottky junction, heterojunction, phase junction, etc. Besides, Ga2O3 avalanche photodiodes, offering ultra-high gain and responsivity, are focused as a promising prototype for commercialization. Furthermore, field effect phototransistors, based on which the scalability and low power performance of Ga2O3 photodetectors have been well proven, are analyzed in detail. Moreover, auxiliary-field configurations with extra tunable dimensions for Ga2O3 photodetectors are introduced. Finally, we conclude this review and discuss the main challenges of Ga2O3 DUV photodetectors from our perspective.

In this letter, the self-powered solar-blind photodetector based on amorphous (a-) SnOx/crystalline (c-) Ga2O3 pn heterojunction with swift response speed has been successfully demonstrated. Increased hole concentration in a-SnOx by enhanced Sn2+ component under annealing improves the self-powered photoresponse performance. A highly precise test scheme without external supply captures the intrinsically swift response speed. The photodetector exhibits an impressive rise/decay time of 0.18/1.04 ms and responsivity of 7.08 A/W at 0 V bias, ranking into the top tier in all self-powered Ga2O3 photodetectors. This study provides a referenceable self-powered detector structure with amorphous/crystalline pn heterojunction.

Low-overlap registration is an important subtask in point cloud registration. In this letter, we focus on two extreme states of low-overlap registration tasks: zero overlap rate and even negative overlap rate point cloud registration. Instead of filtering out non-overlapping regions, overlapping regions are created before registration to deal with the registration tasks with no overlapping regions. Specifically, a novel generative network called Regiffusion is developed on the basis of a diffusion model to predict the shape of the object after registration based on the shapes of the source and target point clouds; the complete object shape, including those of the source point cloud, is accurately predicted before registration and treated as the new target point cloud. This approach effectively creates overlapping regions between the source and target point clouds. We evaluate the developed method on the self-constructed zero-overlap dataset Pokemon-Zero, negative overlap dataset Pokemon-Neg, and the publicly available dataset ModelNet40, indoor datasets 3DMatch. Experimental results demonstrate that the presented method not only performs very welll on zero-overlap and negative overlap datasets, but also improves the registration performance on low-overlap datasets.

Self-powered solar-blind photodetector (SBPD) promises potential applications that urgently need portability and low-power consumption. Herein, an ultrasensitive self-powered p-n heterojunction SBPD based on amorphous NiO and single crystal Ga2O3 has been reliably achieved. The device exhibits a high photo-to-dark-current ratio of ${3}\times {10}^{{6}}$ , ultrahigh responsivity ( ${R}$ ) of 5 A/W, and specific detectivity of ${1.6}\times {10}^{{14}}$ Jones under 254 nm illumination at 0 V, with a solar-blind/visible rejection ratio ( ${R}_{\text {254 nm}}/{R}_{\text {460 nm}}$ ) of ${2}\times {10}^{{4}}$ . Notably, the open circuit voltage can reach 1.3 V and the response speed is significantly less than 1 ms. The comprehensive performance of the device exceeds most reported state-of-the-art Ga2O3 self-powered SBPDs, which is mainly attributed to amorphous NiO/crystalline Ga2O3 vertical junction structure with low-defect interface and strong built-in electric field. This work provides novel design strategies for the future development of high-performance self-powered photodetector.

High tunability of photoresponse characteristics under work conditions is desired for a single solar-blind photodetector to be applied in multifarious fields. Three-terminal metal–oxide–semiconductor field-effect phototransistors have shown excellent controllability of performance, but the hysteresis issue impedes their stable operation. In this work, the metal–semiconductor field-effect phototransistor based on the exfoliated Ga2O3 microflake and graphene thin film is demonstrated. The high-quality quasi-van der Waals interface between Ga2O3 and graphene eliminates the hysteresis issue and generates a subthreshold swing as low as 69.4 mV/dec. By regulating gate voltage (Vg), the dominated mechanism of photocurrent generation in the device can be tuned continuously from the fast photoconduction effect to photogating effect with high photogain. Accordingly, the responsivity, dark current, detectivity, rejection ratio, and decay time of the device can be well adjusted by the Vg. At Vg = −1 V and a source to drain voltage of 2 V, the device shows excellent performance with a responsivity of 2.82 × 103 A/W, a rejection ratio of 5.88 × 105, and a detectivity of 2.67 × 1015 Jones under 254 nm illumination. This work shows the possibility of realizing highly tunable solar-blind photodetectors to meet the requirements for different application fields by introducing gate voltage modulation.

Point cloud completion aims to infer the complete point clouds from incomplete ones, which is used in remote sensing applications such as reconstructing and autonomous driving. However, most existing methods cannot recover accurate structure details of the object. In this article, we propose point shift network (PS-Net). Our main contributions lie in the following three-folds. First, we propose a multiresolution encoder, which extracts and fuses multiresolution point cloud features hierarchically, thus avoiding information loss caused by a single global feature. Second, we design a multiresolution point cloud generation structure, which can be combined with the multiresolution encoder to generate gradually dense point clouds, avoiding the problem of nonuniformly density of the single-layer decoder. Third, we design the shift network (SN), which is used to generate shift vectors to shift the coordinates of each point cloud, so as to further fine-tune the coordinate positions of point clouds, achieving more accurate prediction. We conduct comprehensive experiments on the ShapeNet, KITTI, ScanObjectNN, and ModelNet40 datasets, which demonstrate that the proposed PS-Net achieves better performance than the existing methods and verify the robustness of the proposed method. This article contributes a new method to point cloud completion, realizes fine point cloud shape completion, and brings new possibilities to the research of autonomous driving, registration, and reconstruction.

The dilemma between responsivity and response speed (RS dilemma) limits the performance of photodetectors (PDs). Here, opposite photogating (OPG) engineering is proposed to alleviate this RS dilemma via an auxiliary light source beyond the chip. Based on a $\mathrm{W}\mathrm{S}\mathrm{e}_{2}/\mathrm{G}\mathrm{a}_{2}\mathrm{O}_{3}$ OPG JFET, $\mathrm{a}\gt 10^{3}$ times faster response speed for deep ultra-violet detection has been achieved with negligible sacrifice of responsivity (R). It further demonstrates a R of $2.28\times 10^{3}\mathrm{A}/\mathrm{W}$, specific detectivity of $3.64\times 10^{17}$ Jones, photo-to-dark current ratio of $8.99\times 10^{7}, \mathrm{R}_{\mathrm{D}}\mathrm{u}\mathrm{v}/\mathrm{R}_{\mathrm{v}\mathrm{i}\mathrm{s}\mathrm{i}\mathrm{b}}\mathrm{i}_{\mathrm{e}}$ rejection ratio $\gt 9.6\times 10^{5}$, and external quantum efficiency of $1.11\times 10^{6}$ %. The strong depletion effect on the channel induced by the auxiliary light facilitates the decay speed, while its weak compensation effect sacrifices negligible R. The OPG engineering provides potential inspiration for the design of highly sensitive PDs.