Computers & Graphics-Uk最新文献

In this study, we present SingVisio, an interactive visual analysis system that aims to explain the diffusion model used in singing voice conversion. SingVisio provides a visual display of the generation process in diffusion models, showcasing the step-by-step denoising of the noisy spectrum and its transformation into a clean spectrum that captures the desired singer’s timbre. The system also facilitates side-by-side comparisons of different conditions, such as source content, melody, and target timbre, highlighting the impact of these conditions on the diffusion generation process and resulting conversions. Through comparative and comprehensive evaluations, SingVisio demonstrates its effectiveness in terms of system design, functionality, explainability, and user-friendliness. It offers users of various backgrounds valuable learning experiences and insights into the diffusion model for singing voice conversion.

Understanding the packing of long DNA strands into chromatin is one of the ultimate challenges in genomic research. An intrinsic part of this complex problem is studying the chromatin’s spatial structure. Biologists reconstruct 3D models of chromatin from experimental data, yet the exploration and analysis of such 3D structures is limited in existing genomic data visualization tools. To improve this situation, we investigated the current options of immersive methods and designed a prototypical VR visualization tool for 3D chromatin models that leverages virtual reality to deal with the spatial data. We showcase the tool in three primary use cases. First, we provide an overall 3D shape overview of the chromatin to facilitate the identification of regions of interest and the selection for further investigation. Second, we include the option to export the selected regions and elements in the BED format, which can be loaded into common analytical tools. Third, we integrate epigenetic modification data along the sequence that influence gene expression, either as in-world 2D charts or overlaid on the 3D structure itself. We developed our application in collaboration with two domain experts and gathered insights from two informal studies with five other experts.

Aortic dissection is a life-threatening cardiovascular disease constituted by the delamination of the aortic wall. Due to the weakened structure of the false lumen, the aorta often dilates over time, which can – after certain diameter thresholds are reached – increase the risk of fatal aortic rupture. The identification of patients with a high risk of late adverse events is an ongoing clinical challenge, further complicated by the complex dissection anatomy and the wide variety among patients. Moreover, patient-specific risk stratification depends not only on morphological, but also on hemodynamic factors, which can be derived from computer simulations or 4D flow magnetic resonance imaging (MRI). However, comprehensible visualizations that depict the complex anatomical and functional information in a single view are yet to be developed. These visualization tools will assist clinical research and decision-making by facilitating a comprehensive understanding of the aortic state. For that purpose, we identified several visualization tasks and requirements in close collaboration with cardiovascular imaging scientists and radiologists. We displayed true and false lumen hemodynamics using pathlines as well as surface hemodynamics on the dissection flap and the inner vessel wall. Pathlines indicate antegrade and retrograde flow, blood flow through fenestrations, and branch vessel supply. Dissection-specific hemodynamic measures, such as interluminal pressure difference and flap compliance, provide further insight of the blood flow throughout the cardiac cycle. Finally, we evaluated our visualization techniques with cardiothoracic and vascular surgeons in two separate virtual sessions.

We are experiencing a health literacy crisis worldwide, which has alarming effects on individuals’ medical outcomes. This poses the challenge of communicating key information about health conditions and their management in a way that is easily understood by a general audience. In this paper, we propose the use of data-driven storytelling to address this challenge, in particular through interactive data comics. We developed an interactive data comic that communicates cancer data. A between-group study with 98 participants was carried out to evaluate the data comic’s ease of understanding and engagement, compared to a text medium that captures the same information. The study reveals that the data comic is perceived to be more engaging, and participants have greater recall and understanding of the data within the story, compared with the text medium.

Skinning, a critical process in animation that defines how bones influence the vertices of a 3D character model, significantly impacts the visual effect in animation production. Traditional methods are time-intensive and skill-dependent, whereas automatic techniques lack in flexibility and quality. Our research introduces EasySkinning, a user-friendly system applicable to complex meshes. This method comprises three key components: rigid weight initialization through Voronoi contraction, precise weight editing via curve tools, and smooth weight solving for reconstructing target deformations. EasySkinning begins by contracting the input mesh inwards to the skeletal bones, which improves vertex-to-bone mappings, particularly in intricate mesh areas. We also design intuitive curve-editing tools, allowing users to define more precise bone influential regions. The final stage employs advanced deformation algorithms for smooth weight solving, crucial for achieving desired animations. Through extensive experiments, we demonstrate that EasySkinning not only simplifies the creation of high-quality skinning weights but also consistently outperforms existing automatic and interactive skinning methods.

Instance segmentation is a form of image detection which has a range of applications, such as object refinement, medical image analysis, and image/video editing, all of which demand a high degree of accuracy. However, this precision is often beyond the reach of what even state-of-the-art, fully automated instance segmentation algorithms can deliver. The performance gap becomes particularly prohibitive for small and complex objects. Practitioners typically resort to fully manual annotation, which can be a laborious process. In order to overcome this problem, we propose a novel approach to enable more precise predictions and generate higher-quality segmentation masks for high-curvature, complex and small-scale objects. Our human-assisted segmentation method, HiSEG, augments the existing Strong Mask R-CNN network to incorporate human-specified partial boundaries. We also present a dataset of hand-drawn partial object boundaries, which we refer to as “human attention maps”. In addition, the Partial Sketch Object Boundaries (PSOB) dataset contains hand-drawn partial object boundaries which represent curvatures of an object’s ground truth mask with several pixels. Through extensive evaluation using the PSOB dataset, we show that HiSEG outperforms state-of-the art methods such as Mask R-CNN, Strong Mask R-CNN, Mask2Former, and Segment Anything, achieving respective increases of +42.0, +34.9, +29.9, and +13.4 points in AP${}_{\text{Mask}}$ metrics for these four models. We hope that our novel approach will set a baseline for future human-aided deep learning models by combining fully automated and interactive instance segmentation architectures.

Volumetric radiance fields have been popular in reconstructing small-scale 3D scenes from multi-view images. With additional constraints such as person correspondences, reconstructing a large 3D scene with multiple persons becomes possible. However, existing methods fail for sparse input views or when person correspondences are unavailable. In such cases, the conventional depth image supervision may be insufficient because it only captures the relative position of each person with respect to the camera center. In this paper, we investigate an alternative approach by supervising the optimization framework with a dense pose prior that represents correspondences between the SMPL model and the input images. The core ideas of our approach consist in exploiting dense pose priors estimated from the input images to perform person segmentation and incorporating such priors into the learning of the radiance field. Our proposed dense pose supervision is view-independent, significantly speeding up computational time and improving 3D reconstruction accuracy, with less floaters and noise. We confirm the advantages of our proposed method with extensive evaluation in a subset of the publicly available CMU Panoptic dataset. When training with only five input views, our proposed method achieves an average improvement of 6.1% in PSNR, 3.5% in SSIM, 17.2% in LPIPS^vgg, 19.3% in LPIPS^alex, and 39.4% in training time.

This study introduces a novel framework for choreographing multi-degree of freedom (MDoF) behaviors in large-scale crowd simulations. The framework integrates multi-objective optimization with spatio-temporal ordering to effectively generate and control diverse MDoF crowd behavior states. We propose a set of evaluation criteria for assessing the aesthetic quality of crowd states and employ multi-objective optimization to produce crowd states that meet these criteria. Additionally, we introduce time offset functions and interpolation progress functions to perform complex and diversified behavior state interpolations. Furthermore, we designed a user-centric interaction module that allows for intuitive and flexible adjustments of crowd behavior states through sketching, spline curves, and other interactive means. Qualitative tests and quantitative experiments on the evaluation criteria demonstrate the effectiveness of this method in generating and controlling MDoF behaviors in crowds. Finally, case studies, including real-world applications in the Opening Ceremony of the 2022 Beijing Winter Olympics, validate the practicality and adaptability of this approach.

Ultrasound imaging, characterized by low contrast, high noise, and interference from surrounding tissues, poses significant challenges in lesion segmentation. To tackle these issues, we introduce an enhanced U-shaped network that incorporates several novel features for precise, automated segmentation. Firstly, our model utilizes a convolution-based self-attention mechanism to establish long-range dependencies in feature maps, crucial for small dataset applications, accompanied by a soft thresholding method for noise reduction. Secondly, we employ multi-sized convolutional kernels to enrich feature processing, coupled with curvature calculations to accentuate edge details via a soft-attention approach. Thirdly, an advanced skip connection strategy is implemented in the UNet architecture, integrating information entropy to assess and utilize texture-rich channels, thereby improving semantic detail in the encoder before merging with decoder outputs. We validated our approach using a newly curated dataset, VPUSI (Vascular Plaques Ultrasound Images), alongside the established datasets, BUSI, TN3K and DDTI. Comparative experiments on these datasets show that our model outperforms existing state-of-the-art techniques in segmentation accuracy.

The ShapeBench evaluation methodology is proposed as an extension to the popular Area Under Precision-Recall Curve (PRC/AUC) for measuring the matching performance of local 3D shape descriptors. It is observed that the PRC inadequately accounts for other similar surfaces in the same or different objects when determining whether a candidate match is a true positive. The novel Descriptor Distance Index (DDI) metric is introduced to address this limitation. In contrast to previous evaluation methodologies, which identify entire objects in a given scene, the DDI metric measures descriptor performance by analysing point-to-point distances. The ShapeBench methodology is also more scalable than previous approaches, by using procedural generation. The benchmark is used to evaluate both old and new descriptors. The results produced by the implementation of the benchmark are fully replicable, and are made publicly available.