Stereo matching networks can suffer from generalization challenges when trained on synthetic data and deployed in real-world settings. While existing methods rely on fine-tuning or pre-trained vision foundation models for cross-domain robustness, we revisit this gap from a training perspective and explore a structure-grounded training design that directly improves generalization of RNN-based stereo matching models using only a limited amount of synthetic stereo data, without changing the network architecture or adding any inference overhead. Specifically, we target all three main modules of a typical stereo matching pipeline: in cost volume construction, we enhance geometric cues through data augmentation; in context encoding, we strengthen semantic guidance via auxiliary multitask context supervision; in recurrent disparity refinement, we regulate update dynamics with depth-update regularization. Experiments on multiple mainstream architectures and diverse real-world datasets suggest consistent gains in robustness, improving RAFT-Stereo by 6.6% on KITTI 2015, IGEV-Stereo by 13.7 % on Middlebury, and DLNR by 55.4 % on ETH3D. These insights reveal the previously overlooked importance of structure-grounded training design for achieving reliable stereo depth estimation under data-scarce, domain-shifted conditions.

Simulating event streams from 3D scenes has become a common practice in event-based vision research, as it meets the demand for large-scale, high temporal frequency data without setting up expensive hardware devices or undertaking extensive data collections. Yet existing methods in this direction typically work with noiseless RGB frames that are costly to render, and therefore their simulations are often unrealistically low in temporal resolution. In this work, we propose EventTracer, a path tracing-based rendering pipeline that simulates high-fidelity event sequences from complex 3D scenes in an efficient and physics-aware manner. Specifically, we speed up the rendering process via low sample-per-pixel (SPP) path tracing, and train a lightweight event spiking network to denoise the resulting RGB videos into realistic event sequences. Our EventTracerpipeline runs at a speed of $\sim$∼1 minutes per second of 360p video, and it inherits the merit of accurate spatiotemporal modeling from its path tracing backbone. We show through the Real2Sim and Sim2Real tests that EventTracercaptures higher-fidelity scene details and demonstrates a greater similarity to real-world event data than alternative event simulators, which establishes it as a potential tool for creating large-scale event-RGB datasets, narrowing the sim-to-real gap in event-based vision, and boosting various downstream applications.

High-fidelity 3D video reconstruction is essential for enabling real-time rendering of dynamic scenes with realistic motion in VR/AR. The deformation field paradigm of 3D Gaussian splatting has achieved near-photorealistic results in video reconstruction due to the great representation capability of deep deformation networks. However, in videos with complex motion and significant scale variations, deformation networks often overfit to irregular Gaussian trajectories, leading to suboptimal visual quality. Moreover, the gradient-based densification strategy designed for static scene reconstruction proves inadequate to address the absence of dynamic content. In light of these challenges, we propose a flow-empowered velocity field modeling scheme tailored for Gaussian video reconstruction, dubbed FlowGaussian-VR. It consists of two core components: a velocity field rendering (VFR) pipeline which enables optical flow-based optimization, and a flow-assisted adaptive densification (FAD) strategy that adjusts the number and size of Gaussians in dynamic regions. We also explore a temporal velocity refinement (TVR) post-processing algorithm to further estimate and correct noise in Gaussian trajectories via extended Kalman filtering. We validate our model's effectiveness on multi-view dynamic reconstruction and novel view synthesis with real-world datasets containing challenging motion scenarios, demonstrating not only notable visual improvements (over 2.5 dB gain in PSNR) and less blurry artifacts in dynamic textures, but also regularized and trackable per-Gaussian trajectories.

Abstract RGBD imaging can enable a wide range of applications in computer vision, such as 3D reconstruction, object detection, and scene understanding. While depth estimation through stereo matching has been widely studied, real-world scenarios introduce challenges like defocus blur from the limited depth of field (DoF) in optical systems. Monocular deblurring methods struggle with accuracy, while multi-focus techniques require multiple images, making either impractical for dynamic scenes. These challenges degrade overall RGBD image quality in practical settings. This work presents a learning-based approach that efficiently utilizes the rich information in captured images by an asymmetrically focused stereo camera setup, exploring the complementary relationship between all-in-focus (AiF) image recovery and stereo matching. To realize this, we introduce an iterative error-aware parallel stereo matching (EAP-Stereo) network to improve the robustness of depth estimation under complex real-world conditions, a filter adaptive deblurring and fusion (FADF) network that fully leverages depth information for AiF image recovery, and a SynergyFlow framework that integrates these components through alternating inference. This approach enables effective RGBD imaging, overcoming DoF limitations inherent in single stereo capture. Experimentally, our framework shows superior performance in both synthetic datasets and practical applications, highlighting the potential for seamless and synergistic integration of depth estimation and AiF image recovery.

In the big data era, the healthcare sector grapples with increased data volumes and the push for more intelligent medical devices. This challenge is marked by data silos, higher data processing needs, and the quest for personalized medicine, emphasizing the importance of integrating and analyzing diverse big data from medical devices. This study employs bibliometric analysis to thoroughly analyze the research trends in big data and medical device research, aiming to identify key developments, patterns, and potential future directions in the field. This study employed the Web of Science Core Collection, BIOSIS Citation Index, and Derwent Innovations Index for conducting keyword searches on ’big data’ and ’medical devices’, focusing on English-language articles, reviews, and patents, while excluding duplicates. Quantitative analyses and visualizations were facilitated using tools such as R 4.3.1, VOSviewer, CiteSpace, and Tableau. The research assessed the impact of journals and academic contributions through metrics like the impact factor and G-Index. Our analysis covered 592 articles and 795 patents. Notably, the annual growth rate of articles reached 62.19%, with the primary contributions originating from China, the United States, India, and England. Among the identified publications, “IEEE Access” emerged as the most prominent journal. The research identified key trends in the application of big data within medical devices, including the extensive use of artificial intelligence (AI), deep learning, Internet of Things (IoT) technologies, genomic data analysis, and bioinformatics. Case studies show how big data and medical device integration, exemplified by Electronic Health Records (EHRs), Clinical Decision Support Systems (CDSS), and smart wearables, enhance diagnostic efficiency and care quality. The study further identified intensive care units, radiology departments, and oncology as key application departments leveraging this integration. Big data and medical device integration have led to smart healthcare and personalized treatment, improving medical decision accuracy and inspiring new research directions. Future research efforts should explore the potential for incorporating even more expansive datasets and multilingual resources to further maximize the potential of this integration and advance medical technology in disease prevention, diagnosis, treatment, and management.