The proceedings contain 76 papers. The topics discussed include: detecting ambiguities in 3D polygons with self-intersecting projections;an example-driven symbol recognition approach based on key features in engineeri...
ISBN:
(纸本)9780769544977
The proceedings contain 76 papers. The topics discussed include: detecting ambiguities in 3D polygons with self-intersecting projections;an example-driven symbol recognition approach based on key features in engineering drawings;niche improved particle swarm optimization on geometric constraint solving;importance-driven composition of multiple rendering styles;multi-GPU compositeless parallel rendering algorithm;structure preserving mesh parameterization;real-time image smoothing based on true edges;color-mood-aware clothing re-texturing;automatic segmentation of head-and-shoulder images by combining edge feature and shape prior;creating cylindrical panoramic mosaic from a pipeline video;detection geometric object in the conformal geometric algebra framework;a unified framework for alpha matting;variational harmonic method for parameterization of computational domain in 2D isogeometric analysis;and convergence of geometric interpolation using uniform B-splines.
Homomorphic encryption enables computations on the ciphertext to preserve data privacy. However, its practical deployment has been hindered by the significant computational overhead compared to the plaintext computati...
详细信息
ISBN:
(纸本)9798400710773
Homomorphic encryption enables computations on the ciphertext to preserve data privacy. However, its practical deployment has been hindered by the significant computational overhead compared to the plaintext computations. In response to this challenge, we present HERMES, a novel hardware acceleration system designed to explore the computation flow of the CKKS homomorphic encryption bootstrapping process. Among the major contributions of our proposed architecture, we first analyze the properties of the CKKS computation data flow and propose a new scheduling strategy by partitioning the computation modules into general-purpose and special-purpose modular computation modules to allow smaller resource consumption and flexible scheduling. The computation modules are also reconfigurable to reduce the memory access overhead during the intermediate computation. We also optimize the CKKS computation dataflow to improve the regularity with reduced control overhead.
This paper proposes to integrate reconfigurable intelligent surfaces (RISs) with unmanned aerial vehicles (UAVs) as a resilience mechanism to mitigate outages in UAV networks due to UAV and link failures. The inherent...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
This paper proposes to integrate reconfigurable intelligent surfaces (RISs) with unmanned aerial vehicles (UAVs) as a resilience mechanism to mitigate outages in UAV networks due to UAV and link failures. The inherent addition of RIS-aided links (UE-RIS-UAV links), combined with their reconfigurability, creates alternative paths for user equipment (UEs) to transmit signals to UAVs. The paper studies the problem of maximizing connectivity of UAV networks by jointly considering UE positioning, RIS-aided link selection, and phase shift design of RISs. To tackle it, we propose an efficient two-step solution. In the first step, we propose a supergradient method that locates the UEs in positions that improve their communication links until a certain connectivity threshold is satisfied. Given the optimized UE positioning, the second step jointly optimizes the RIS-aided link selection and RIS phase shift design using semidefinite programming (SDP). Through simulations, we illustrate the superiority of the proposed solution compared to the solutions available in the literature.
The work presents a novel way to exhibit art by integrating mixed reality (MR) technology for displaying art. Conventional display spaces are often limited in terms of space and access. We created and developed a digi...
详细信息
ISBN:
(数字)9798331521578
ISBN:
(纸本)9798331521585
The work presents a novel way to exhibit art by integrating mixed reality (MR) technology for displaying art. Conventional display spaces are often limited in terms of space and access. We created and developed a digital art display system that utilizes MR elements to enable users to see the artworks in mixed-reality environments and interact with them using MR interfaces. The system encompasses 3D scanning, real-time rendering, and MR hardware to produce an immersive and interactive experience for viewing art. Our methodology involved system design, optimization, technical implementation, user interface design and user studies. We conducted experiments to determine user experiences and engagement with the developed system. Results indicated higher user engagement and immersiveness levels with minimized lag in the MR-enhanced virtual display. Our system has intuitive MR interfaces for artwork interaction. Optimization techniques are proposed to overcome technical challenges such as low latency. The study shows how the proposed work can be the future of art exhibitions by improving access, interactivity, and participation.
Breast cancer still ranks among the most common causes of cancer-related deaths among women, hence the call for early diagnosis. Mammography is the most accepted screening test, but conventional computer-aided detecti...
详细信息
ISBN:
(数字)9798331521691
ISBN:
(纸本)9798331521707
Breast cancer still ranks among the most common causes of cancer-related deaths among women, hence the call for early diagnosis. Mammography is the most accepted screening test, but conventional computer-aided detection (CAD) has a high false positive rate (FPR) that gives rise to biopsy and false negatives (FN) where cancer is undetected. In solving these challenges, this paper provides a solution by employing the use of the Simple Genetic Algorithm (SGA), which is openly inspired by biological systems to enhance the performance of CAD systems for breast cancer detection. The SGA, which is based on the evolutionary process, can resolve problems in feature selection and classification o f t he m ammogram b y overcoming shortcomings of pattern recognition. By mimicking the genetic evolution process, ant colony optimization, and swarm intelligence, the SGA prevents noisy or variant images from anyhow decreasing the detection accuracy. Comprehensive tests on typical sets of mammograms confirm t he e ffectiveness o f t he proposed approach regarding a twofold reduction of inappropriate positive and negative results. This enhanced accuracy of diagnoses can help radiologists to act early, combined with favorable outcomes for the patients, implying that early diagnosis may save lives.
The proceedings contain 82 papers. The topics discussed include: a robust iterative method devoted to pole curve fitting;evolution of 3D curves under strict spatial constraints;a fitting approach with dynamic algebrai...
详细信息
ISBN:
(纸本)0769524737
The proceedings contain 82 papers. The topics discussed include: a robust iterative method devoted to pole curve fitting;evolution of 3D curves under strict spatial constraints;a fitting approach with dynamic algebraic spline curves;a method of generating smooth surface using C-curves;determination of free parameters in algebraic surface blending;surface blending with curvature continuity;the constraint elevation for parametric continuity order of NURBS curve and surface;subdivision surfaces based on point-based splines;adaptive refinement scheme for subdivision surfaces based on triangular meshes;high quality triangulation of implicit surfaces;implicit fitting and smoothing using radial basis functions with partition of unity;fast mesh-free deformations;3D model retrieval based on orthogonal projections;and rod-like trabeculae extraction from cancellous bone microstructure using topological analysis.
The proceedings contain 47 papers. The topics discussed include: an adaptive spherical collision detection and resolution method for deformable object simulation;automatic hierarchical mid-surface abstraction of thin-...
ISBN:
(纸本)9781467380201
The proceedings contain 47 papers. The topics discussed include: an adaptive spherical collision detection and resolution method for deformable object simulation;automatic hierarchical mid-surface abstraction of thin-walled models based on rib decomposition;hierarchical mesh segmentation based on quadric surface fitting;efficient depth restoration from 2D line drawings with line segments and curves;a piecewise cubic polynomial interpolation algorithm for approximating elementary function;construction of m-ary symmetric interpolatory subdivision scheme by m-ary orthonormal refinable function;generate accurate soft shadows using complete occluder buffer;novel, robust, and efficient guidewire modeling for PCI surgery simulator based on heterogeneous and integrated chain-mails;step into micro-world: dynamic simulation of the coffee-ring effect;an integrated de-noise and enhancement method for ancient Chinese tablet images;and real time learning evaluation based on gaze tracking.
Inverse rendering has been extensively explored with the advent of neural implicit fields. However, existing methods struggle to model global illumination and to fully integrate volume rendering with physically based ...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
Inverse rendering has been extensively explored with the advent of neural implicit fields. However, existing methods struggle to model global illumination and to fully integrate volume rendering with physically based rendering in a single stage. To address this issue, we propose TIRPL, a tailored-made inverse rendering method for point-light scenes. Our method efficiently combines volume rendering and physically based rendering to jointly optimize scene geometry, materials, and global illumination from multi-view images in a single stage. We design an indirect color network and allow the light intensity to be learnable to estimate global illumination. Additionally, a feature vector and a tailored-made light intensity decay rate are devised to combine volume rendering and physically based rendering in a single stage. Extensive experiments are conducted in both real and synthetic datasets, showing the effectiveness of our method.
Peer-to-peer learning is an increasingly popular framework that enables beyond-5G distributed edge devices to collaboratively train deep neural networks in a privacy-preserving manner without the aid of a central serv...
详细信息
ISBN:
(数字)9798350368741
ISBN:
(纸本)9798350368758
Peer-to-peer learning is an increasingly popular framework that enables beyond-5G distributed edge devices to collaboratively train deep neural networks in a privacy-preserving manner without the aid of a central server. Neural network training algorithms for emerging environments, e.g., smart cities, have many design considerations that are difficult to tune in deployment settings – such as neural network architectures and hyperparameters. This presents a critical need for characterizing the training dynamics of distributed optimization algorithms used to train highly nonconvex neural networks in peer-to-peer learning environments. In this work, we provide an explicit characterization of the learning dynamics of wide neural networks trained using popular distributed gradient descent (DGD) algorithms. Our results leverage both recent advancements in neural tangent kernel (NTK) theory and extensive previous work on distributed learning and consensus. We validate our analytical results by accurately predicting the parameter and error dynamics of wide neural networks trained for classification tasks.
This paper proposes a research method of urban environment design and renewal simulation model based on virtual reality technology, aiming to improve the efficiency and accuracy of urban planning and renewal process. ...
详细信息
ISBN:
(数字)9798350389579
ISBN:
(纸本)9798350389586
This paper proposes a research method of urban environment design and renewal simulation model based on virtual reality technology, aiming to improve the efficiency and accuracy of urban planning and renewal process. This method takes point cloud data processing as the core technology, combines advanced 3D modeling and virtual reality technology, and constructs a high-precision urban environment simulation system. The system includes key modules such as data acquisition, optimization algorithm design, model construction, simulation verification and interactive operation. Among them, through the improved point cloud data processing algorithm, the rapid reconstruction of complex urban spatial structure is realized, and efficient 3D rendering and large-scale scene simulation are supported. Data analysis of multiple actual urban scenes shows that this method has excellent performance in various environments. The simulation system supports accurate visual comparison of different urban design schemes, interactive scene adjustment through real-time dynamic operation, and quantitative evaluation of design results. Experimental data show that the system can control the simulation error within 5% in multiple urban renewal scenarios; the data processing efficiency is improved by about 30% compared with traditional methods, and the 3D reconstruction time is shortened to 70% of the original.
暂无评论