A multiscale Canny edge detection is equivalent to finding the local maxima of a wavelet transform. The authors study properties edges through theory. For pattern recognition, one often needs discriminate different types edges. They show that evolution across scales characterize shape irregular structures. Numerical descriptors are derived. completeness representation also studied. describe an algorithm reconstructs close approximation 1-D and 2-D signals from their images, reconstruction...

For software quality estimation, development practitioners typically construct quality-classification or fault prediction models using metrics and data from a previous system release similar project. Engineers then use these to predict the proneness of modules in development. Software estimation supervised-learning approaches is difficult without measurement projects earlier releases. Cluster analysis with expert input viable unsupervised-learning solution for predicting modules' potential...

Line segment extraction from point clouds is critical for analyzing and understanding large-scale scenes. The main challenge to generate line segments accurately as well completely. However, state-of-the-art approaches continue struggle with this issue. In paper, we propose a novel method effectively generating clouds. To end, design weighted centroid displacement scheme identifying comprehensive feature points. Then, employ an L1-median optimization refine the identified points perceive...

A variety of Coupled HMMs (CHMMs) have recently been proposed as extensions HMM to better characterize multiple interdependent sequences. This paper introduces a novel distance coupled HMM. It then compares the performance several and CHMM models for multi-channel EEG classification problem. The results show that, all approaches examined, multivariate that has low computational complexity surprisingly outperforms other models.

Localization in the presence of malicious beacon nodes is an important problem wireless networks. Although significant progress has been made on this problem, some fundamental theoretical questions still remain unanswered: nodes, what are necessary and sufficient conditions to guarantee a bounded error during 2-dimensional location estimation? Under these conditions, class localization algorithms can provide that bound? In paper, we try answer questions. Specifically, show that, when number...

High-fidelity garment reconstruction is essential for various applications such as design and virtual try-on. While image-based methods have made significant progress with deep generative models, generating 3D models from hand-drawn sketches to meet intentions remains challenging. One of the main obstacles limited availability large-scale accompanied by corresponding sketches. To address this issue, we propose SGM, a comprehensive dataset comprising 656 categorized into short long sleeves....

Speckle reduction is a fundamental problem in coherent imaging systems. In this paper, to suppress the speckle SAR images, we propose novel truncated nonconvex nonsmooth model. It incorporates regularization term and an I-divergence fidelity term. The ℓpnorm (0<p<1) can better recover neat edges simultaneously prevent contrast artefact. used multiplicative noise effectively. We also efficient algorithm based on variable-splitting alternating direction method of multipliers (ADMM) solve...

Abstract Mesh denoising is a fundamental problem in geometry processing. The main challenge to preserve sharp features (such as edges and corners) smooth regions smoothly curved fine details) while removing the noise. State‐of‐the‐art methods still struggle with this issue. In paper, we first propose new variational model combining total variation anisotropic Laplacian regularization filter normal vector field of mesh. This can simultaneously handle well. Then, vertex updating scheme...

An algorithm that reconstructs one-dimensional signals and images from their sharper variation points at dyadic scales is described. This exactly multiscale edges. It proved the evolution across of wavelet maxima characterizes local shape sharp variations signal. One can thus not only detect edges but also classify them. The representation a new reorganization image information makes it possible to develop algorithms uniquely based on for solving processing problems.< <ETX...

The hybrid artificial bee colony (ABC) algorithm with differential evolution (DE) techniques (HABCwDE) is proposed for numerical optimization in this paper. HABCwDE adopts multiple candidate solution generation strategies (CSGSes) from DE to generate new solutions the framework of ABC algorithm. In algorithm, three CSGSes and groups parameter settings are employed. performance some other evolutionary algorithms tested on 26 state-of-the-art benchmark functions. Experimental results...

The edges of an image can be detected at different scales from the local maxima its wavelet transform. algorithm is described that reconstructs images their dyadic scales. representation a novel reorganization information makes it possible to develop algorithms uniquely based on for solving processing and computer vision problems. evolution across gives precise characterization edge type which used pattern recognition. A coding selects most important in order obtain compact representation.<...

Mesh denoising is to recover high quality meshes from noisy inputs scanned the real world. It a crucial step in geometry processing, computer vision, computer-aided design, etc. Yet, state-of-the-art methods still fall short of handling containing both sharp features and fine details. Besides, some usually introduce undesired staircase effects smoothly curved regions. These issues become more severe when mesh corrupted by various kinds noise, including Gaussian, impulsive, mixed...

A software quality estimation model is often built using known metrics and fault data obtained from program modules of previously developed releases or similar projects. Such a supervised learning approach to assumes that available for all the modules. Considering various practical issues in project development, may not be training data. More specifically, labeled such yield good prediction. In contrast, classification scheme aided by unlabeled data, i.e., semisupervised learning, better...

In this paper, we propose a novel guided normal filtering followed by vertex updating for mesh denoising. We introduce two-stage scheme to construct adaptive consistent neighborhoods filtering. the first stage, newly design consistency measurement select coarse neighborhood each face in patch-shift manner. step, selected may still contain some features. Then, graph-cut based is iteratively performed constructing different match corresponding local shapes of mesh. The constructed known as...

Rigid registration of 3D indoor scenes is a fundamental yet vital task in various fields that include remote sensing (e.g., reconstruction scenes), photogrammetry measurement, geometry modeling, etc. Nevertheless, state-of-the-art approaches still have defects when dealing with low-quality scene point clouds derived from consumer-grade RGB-D sensors. The major challenge accurately extracting correspondences between pair they contain considerable noise, outliers, or weak texture features. To...

A second generation image coding algorithm based on multiscale edges is described. Edges are detected from the local maxima of wavelet transform. The original can be reconstructed corresponding edge map. For purposes, that important for visualization selected, and an efficient encoding data High compression ratio obtained images where texture information not important. Examples with ratios around 30 given.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML"...

Abstract. Building the fine 3D model from outdoor to indoor is becoming a necessity for protecting cultural tourism resources. However, existing modelling technologies mainly focus on areas. Actually, should contain detailed descriptions of both its appearance and internal structure, including architectural components. In this paper, portable four-camera stereo photographic measurement system developed, which can provide professional solution fast data acquisition, processing, integration,...

3D human reconstruction from a single image has achieved great progress with recent deep neural networks. However, conventional approaches still struggle the issues of over-smoothing details and wrong limb poses. To this end, we propose PMDI, method that combines parametric-model depth-aware implicit function for single-view reconstruction. First, given an RGB person's whole body as input, predicts its corresponding SMPL parameter model, depth map, front (back) normal map by using Then,...