Table of Contents

28 February 2019, Volume 40 Issue 1 Previous Issue    Next Issue

---

For Selected:

Download Citations EndNote Ris BibTeX

Toggle Thumbnails

Select

As-Rigid-As-Possible Mesh Deformation Controlled by  Geometric Field in Heat SHAO Mao-zhen, SHOU Hua-hao 2019, 40(1): 1-7.  DOI: 10.11996/JG.j.2095-302X.2019010001 Abstract ( 201 )   PDF (1444KB) ( 258 )   In order to maintain the details of the 3D model, correct the problem of distortion and folding of the as-rigid-as-possible (ARAP) mesh deformation used in large and nonperfect rigid deformation, an ARAP deformation method is proposed based on geometric field in heat. First, the Laplacian deformation of the model is carried out. On this basis, the rotation matrix of local cell is solved by singular value decomposition, and the rigid deformation energy of the model is calculated. Then by solving the sparse linear system, the deformation points are updated. By solving the two-time sparse linear system, we calculate the geometric field deviation of the deformation process, and correct the deformed mesh to get the deformation results close to those of the original mesh. Iterate the above steps until the geometric field deviation to meet certain requirements, and finally the final deformation results are obtained. The example shows that the method can quickly complete the mesh point correction function in mesh deformation process, and it can also effectively avoid grid collapse when applied to large-scale deformation. Related Articles | Metrics

Select

Trajectory Modeling in Short Track Speed Skating LI Qi1,2, YANG Wen-hao3, HOU Ya-li4, WANG Xiang-dong4, LI Hua1,2 2019, 40(1): 8-14.  DOI: 10.11996/JG.j.2095-302X.2019010008 Abstract ( 168 )   PDF (12236KB) ( 153 )   This paper explores the question of detecting and tracking multiple athletes in a short track speed skating video and reconstructing their trajectories. The current technology is difficult to ensure that each target is completely tracked and its trajectory is recovered in a complex motion video with frequent blocking and interlaced positions. We propose a complete algorithm flow to quickly and accurately detect and track athletes. Further, a trajectory reconstruction algorithm based on cubic B-spline curve is proposed. Using the nonuniformity and continuity of the B-spline curve, this algorithm can completely match the smooth motion trajectory of each target in the absence of partial tracking results. Experiments show that this algorithm can detect and track multiple athletes and rebuild their trajectories, and the results obtained can be further used in the actual technical and tactical analysis. Related Articles | Metrics

Select

Geometric Iteration Method Based on Many-Knot Spline Polishing Functions HUO Yan-wen, CAI Zhan-chuan 2019, 40(1): 15-23.  DOI: 10.11996/JG.j.2095-302X.2019010015 Abstract ( 150 )   PDF (1083KB) ( 251 )   Geometric iteration method has been widely used in computer aided geometric design (CAGD). In order to improve the convergence speed and iterative accuracy of the traditional B-spline curve interpolation in geometric iterations, this study proposes the geometric iteration method based on many-knot spline polishing functions, which introduces many-knot spline polishing functions, and combines many-knot spline polishing functions method and geometric iteration method in curve fitting. After polishing operator and iterating, the curve fitting method with high approximation under the optimal solution of L-BFGS iterative algorithm is constructed. Experimental results show that the proposed method not only reduces the times of iterations, but also improves the iterative speed under the same accuracy. The proposed geometric iteration method can be used in the shape design of airplanes, automobiles, etc. It can also be used to reconstruct and rebuild the shape of cultural relic houses and satellite image processing. Related Articles | Metrics

Select

A Method for Visual Analysis of Urban Power Supply Data Based on  Power Line Layout Optimization LU Qiang1,4, ZHANG Hai-bo1, YUAN Xiao-hui2, CHEN Chen3,4 2019, 40(1): 24-31.  DOI: 10.11996/JG.j.2095-302X.2019010024 Abstract ( 86 )   PDF (2320KB) ( 258 )   The visualization of power data can realize the online monitoring of different power features including various attributes and operating status, etc., which can be graphically and visually presented to provide a powerful guarantee for the timely and effective monitoring and analysis of equipment operation status. Based on the urban power supply data, a method for the visual analysis of power supply situation in the city is proposed. Firstly, the overall power supply situation is visualized based on the substation data in the power supply system. Then, the layout is optimized based on multi-criteria constraints and visually coded according to the power supply line data. Finally, according to the relation of lines, the visualized coding is conducted on the users of different lines. A visual analysis system has also been designed to facilitate the user’s exploration and analysis of the power supply situation in a certain area. Case studies based on Shanghai City Power Supply data show that the method can effectively reflect the overall power supply situation and specific power supply details in a certain area. Related Articles | Metrics

Select

Simulation of Hail Impact on Leaves Based on Mass-Spring Model XIE Jia-yi, YANG Meng,LIANG Wan-ru, CUI Ya-ting, QI Xiao-bin, YANG Gang 2019, 40(1): 32-39.  DOI: 10.11996/JG.j.2095-302X.2019010032 Abstract ( 118 )   PDF (8795KB) ( 75 )   In order to realize the simulation of the process of hail against the leaves, which is different from the fabric crushing simulation, this paper presents a realistic dynamic simulation technology of the interaction between leaves and hail based on physical theory. It opens up a different area from the cloth tearing simulation and proposes the simulation algorithm of the leaf breaking, which provides the foundation for further research on the influence of leaf veins on leave’s force-bearing capability. The tearing of the leaves is due to the hail being attracted by the gravitation of the ground, transferring the energy of the hail to the leaves, and the interaction between the two can be well simulated by an efficient technique. The technology is specifically designed to simulate the tearing of objects with less elastic forces such as leaves. When hail interacts with the leaves, the forces due to the hail velocity are translated into lateral and longitudinal spring forces on the leaves, causing the leaves to break. Experimental results show that this technique can effectively and realistically simulate the effect of hail impact on leaves and restore the situation of hail disasters. Related Articles | Metrics

Select

A Method of 3D Vertebra Segmentation Based on Maximum Feature WANG Hao1, SHAO Kun1, HUO Xing2, YANG Peng1, TAN Jie-qing2 2019, 40(1): 40-45.  DOI: 10.11996/JG.j.2095-302X.2019010040 Abstract ( 237 )   PDF (1419KB) ( 167 )   Accurate segmentation of vertebrae has become more and more important for the diagnosis and treatment of spinal diseases. After studying the change rule of some spinal CT image sequences of normal people, we proposed a new vertebral segmentation algorithm based on vertebral area change regulation. This method first calculates the area of vertebral region in the image after pre-processing, then finds out the significant features for segmentation, uses image similarity of continuous images to select the actual segmentation points, and finally extracts vertebral images from image sequence for 3D reconstruction. The innovation of the method is that a new segmentation algorithm based on the maximum feature and correlation of CT image sequence is proposed. The experimental results show that the algorithm has a good segmentation effect on the CT image sequence of lumbar vertebrae and the lower thoracic vertebrae of normal human bodies and it has a high degree of automation. This study is of great significance to the morphological study of the spine and the simulation of the scoliosis operation. Related Articles | Metrics

Select

Design of Smooth Surfaces with Regular Network of Curvature Curves ZHENG Hui-ying, BO Peng-bo 2019, 40(1): 46-53.  DOI: 10.11996/JG.j.2095-302X.2019010046 Abstract ( 167 )   PDF (9385KB) ( 93 )   A new algorithm is proposed for the design of a smooth surfaces which is required to have a regular mesh of curvature curves without any umbilical points. A composite surface as smoothly joint by Bézier surface patches is used in the design procedure. By the alternative steps of optimization of curvature line networks and surface subdivision, the free from surface is generated which has a regular network of curvature curves without an umbilical point. Based on the resultant surface, a cyclide spline surface with regular structures, conical strip models and multi-layer support structures for freeform architectural surface design are easily obtained. Related Articles | Metrics

Select

Generalized Barycentric Coordinates Based on Combinatorial  Optimization of Sparse Solutions PENG Feng-fu, FANG Ming 2019, 40(1): 54-58.  DOI: 10.11996/JG.j.2095-302X.2019010054 Abstract ( 78 )   PDF (751KB) ( 196 )   According to the nature and characteristic of the linear operation of generalized barycentric coordinates, by means of a combination of weighted harmonic mean funcitons, an algorithm for solving generalized barycentric coordinates is designed to meet the demands of the vertexes of spatial convex polyhedron, in which the weighted function is a unary function with conformal parameters, thus it is characterized with conformal optimization. Two different types of weighted functions are constructed in this paper, and they are both used to calculate the generalized barycentric coordinates. An example of a plane figure is mapped into a space surface by the means, which is to be described and analyzed with different weighted functions and parameters. By means of their contours, the generalized barycentric coordinates for the example are analyzed and compared. Related Articles | Metrics

Select

Construction of Ck Continuous Transition Curve Based on Metaball Technique YAN Lan-lan, FAN Ji-qiu, HUANG Tao 2019, 40(1): 59-69.  DOI: 10.11996/JG.j.2095-302X.2019010059 Abstract ( 87 )   PDF (477KB) ( 126 )    In order to obtain the general expression of polynomial potential function which can make the corresponding transition curve reach Ck ( k is an arbitrary natural number) continuity at the endpoints, based on the required conditions of the potential function deduced from the continuity condition, the minimal times of the polynomial potential function is determined according to the number of the required conditions, and the potential function is expressed as a linear combination of the classical Bernstein basis functions which are yet to be determined. According to the function values and derivative values at the endpoints of the Bernstein basis functions, the required conditions of the potential function are converted into an equation set about the undetermined combination coefficients. Solving the equation, we obtain the general expression of the potential function which satisfies the expected continuity condition. Considering that the shape of the transition curve is uniquely determined by the curves needed to be transferred, another new polynomial potential function with one higher degree is constructed. By similar derivation process, the general expression of the new potential function which can make the transition curve reach arbitrary Ck continuity and with shape adjustability is obtained. With the help of the degree elevation formula of the classical Bernstein basis functions, the relationship between the two kinds of potential functions is deduced. The properties of the potential functions and the characteristics of the corresponding transition curves are also analyzed. Legends of the potential functions and the transition curves are put forward and they can verify the correctness of the theoretical analysis results and the validity of the presented method. Related Articles | Metrics

Select

Tongue Segmentation Algorithm Based on Improved Bacteria Foraging Optimization Algorithm ZHU Feng1, DAI Pu-wei1, PAN Bin1, GUO Xiao-ming1, WANG Yu-ming1, ZHONG Fan2 2019, 40(1): 70-77.  DOI: 10.11996/JG.j.2095-302X.2019010070 Abstract ( 99 )   PDF (1997KB) ( 155 )   Aimed at the accurate segmentation of medical digital image of tongue, this study proposed a tongue segmentation algorithm based on bacteria foraging optimization algorithm (BFOA) and Snake active contour model. Firstly, information entropy threshold method and Kapur algorithm are combined to improve BFOA, and then the improved BFOA is used to get the optimal threshold value, then binarize the tongue image. Secondly, the key contour points of the tongue coating are extracted according to the symmetry. Finally, the B-spline interpolation is performed by using the key points to obtain the closed B-spline curve as the initial contour of the Snake model, by which the external contour of the tongue edge is accurately extracted. The results show that the proposed algorithm can improve the accuracy of tongue segmentation and reduce human-computer interaction. Related Articles | Metrics

Select

FMR: A Fast Multilevel Algorithm Based on FR ZHANG Ye, WANG Song, WU Ya-dong 2019, 40(1): 78-86.  DOI: 10.11996/JG.j.2095-302X.2019010078 Abstract ( 88 )   PDF (5165KB) ( 164 )   Graph visualization technology is an important part of visualization research. The drawing of large graphs has been the focus of graph visualization technology in recent years. This paper proposes a fast multilevel algorithm for solving large graph drawing problems. The algorithm uses a multilevel method as the framework of the algorithm. The present study uses a FR force-directed algorithm variant combined with centroid algorithm and quad-tree space decomposition method to refine the single-level layouts. Also, energy-model and constraint-normalization are used. Experiments show that the algorithm has high performance and good layout effect. The algorithm is of high efficiency, and under a single-core CPU, it can effectively draw a graph of 10,000 vertices in about 5 seconds. Through comparison with several classical algorithms, the effectiveness and practicability of the algorithm was proved. In addition, the algorithm is easy to implement and can be easily generalized to other layout algorithms to speed up its operations. Related Articles | Metrics

Select

A Scene-Distributed Interactive Rendering System SUN Zhao1, LIU You-quan1, ZHANG Cai-rong1, SHI Jian2, CHEN Yan-yun3 2019, 40(1): 87-91.  DOI: 10.11996/JG.j.2095-302X.2019010087 Abstract ( 118 )   PDF (452KB) ( 163 )   With the constant increase in the scale and complexity of three-dimensional virtual scenes in recent years, very large-scale scenes (such as buildings, cities, etc.) can hardly be processed on single rendering node and their interactive performance can also be hard to achieve. Aiming at this shortcoming, the authors present an interactive distributed rendering framework. The authors divide the large scene into a set of renderable sub-scenes, which are distributed to different rendering nodes for processing. Intermediate sub-scene rendering results are merged to the final result based on depth information. To reduce the latency, the rendering results are compressed to accelerate the network transmission. The proposed distributed rendering system can efficiently process rendering and interaction for large-scale scenes. Moreover, the experiment has also confirmed that it is able to provide good scalability. A wide range of applications can benefit from interactive distributed rendering by this system. Related Articles | Metrics

Select

Progressive Transmission Method Based on Dihedral Reverse  Interpolation Loop Subdivision SHI Zhuo, KONG Qian, YU Ke, LAN Ru-shi, LUO Xiao-nan 2019, 40(1): 92-98.  DOI: 10.11996/JG.j.2095-302X.2019010092 Abstract ( 101 )   PDF (1220KB) ( 150 )   With the rapid development of virtual reality, augmented reality, etc., Multi-resolution progressive transmission can provide a good user experience. In order to implement the fast transmission and display of triangular mesh in the mobile terminal, this paper presents a progressive transmission algorithm based on dihedral reverse interpolation Loop subdivision (DRILS). The algorithm mainly uses dihedral interpolation Loop subdivision (DILS) and interpolation Loop subdivision (ILS) algorithm processing. A fine mesh with subdivision connectivity is obtained at the same time as the local feature is accurate. In the process of progressive transmission, three steps are taken for the iterative operation of the accurate triangular mesh: odd and even vertex partition, offset prediction, and triangular mesh prediction. Due to the combination of the DILS and ILS which is used to obtain accurate mesh, the local characteristics of the progressive transmission are accurate, and the speed of progressive transmission is also accelerated. The experimental comparison shows that the algorithm is accurate and efficient, and is suitable for display, transmission and storage of mobile terminal devices. Related Articles | Metrics

Select

A Fast Water Information Extraction Method Based on  GF-2 Remote Sensing Image ZOU Cheng1,2, YANG Xue-zhi1,2, DONG Zhang-yu1,2, WANG Dong1,2 2019, 40(1): 99-104.  DOI: 10.11996/JG.j.2095-302X.2019010099 Abstract ( 198 )   PDF (1274KB) ( 294 )   It is difficult to distinguish water from shadow (especially the shadows of tall buildings) and dark ground objects in high resolution remote sensing images, especially in GF-2 remote sensing images. This study analyzes the spectral features of typical terrains of the GF-2 remote sensing images through a lot of experiments. A new comprehensive water index method (NCWI) is proposed to enhance water body region information; and then, the improved method of maximum between-class variance (OSTU) combining with the chicken swarm optimization algorithm (CSO) are used to quickly and adaptively determine the optimal segmentation threshold to obtain the final water body region. To demonstrate the effectiveness of the proposed algorithm, the method of NDWI algorithm, the multi-band spectrum-photometric algorithm and the principal component analysis synthesis algorithms are used for comparison in water-body extraction. The confusion matrix and the field sampling are applied as the statistical metric to quantitatively evaluate the performance of the algorithms mentioned above. The verification results indicate that the new method can be used to extract quickly and effectively extract water body information, and the accuracy reached 97.82%, 97.44%, 92.13%, 96.94% respectively. Related Articles | Metrics

Select

Parameter-Free Uncorrelated Maximum Discriminant Margin Algorithm LIANG Xing-zhu, LIN Yu-e, XU Guang-yu 2019, 40(1): 105-110.  DOI: 10.11996/JG.j.2095-302X.2019010105 Abstract ( 104 )   PDF (247KB) ( 118 )   The parameter-free locality preserving projection (PFLPP) is an effective feature extraction algorithm for face recognition, but it cannot effectively determine the neighbor relationship because it does not consider neighborhood relationship between the samples from different classes, and this algorithm judges the neighborhood relationship only by the distance between the samples and the population mean. In this paper, parameter-free uncorrelated maximum discriminant margin algorithm is proposed, which effectively uses the class information of the samples and needn’t set any parameters. The algorithm defines the similar weights of the neighbor samples from the same class and the punishment weights of the neighbor samples from different classes. The size of the sample neighborhood can be adaptively determined by the mean of the intraclass cosine distance and the inter-class cosine distance. In order to further enhance the performance of the algorithm, the uncorrelated objective function based on the maximum discriminant margin is put forward. The experimental results of UMIST and AR face database show that the proposed method has the advantages of low computation and high recognition performance compared with PFLPP and uncorrelated locality preserving projections analysis. Related Articles | Metrics

Select

Image Denoising Algorithm Based on Noise Detection and Dynamic Window WANG Wen-hao, YAN Yun-yang, JIANG Ming-xin, GAO Shang-bing, YU Yong-tao 2019, 40(1): 111-116.  DOI: 10.11996/JG.j.2095-302X.2019010111 Abstract ( 89 )   PDF (4753KB) ( 232 )   Aimed at the loss of image details caused by median filtering when removing the impulse noise, this study proposes an adaptive filtering method based on noise detection and dynamic window. Firstly, using BDND method, the image pixel is divided into signal points and suspected noise points to reduce the number of pixels that need to be processed. Then a method of the window adaptive noise detection is designed to further distinguish the suspected noise points into the noise points and the fine points, which strengthen the protection of the details of the image. Finally the detected noise is removed by an improved adaptive median filter. The window adaptive control is integrated into this filter algorithm. The size of the window can be adaptively adjusted according to the noise condition. The image details are protected as much as possible while removing the noise. The experiments show that the algorithm in this paper is superior to other conventional algorithms in noise removal and detail protection, and can effectively improve the peak signal to noise ratio of the image. This method can also obtain better denoising effect for images with high density noise. Related Articles | Metrics

Select

Remote Sensing Image Change Detection Based on  Relief-PCA Feature Selection WANG Shou-feng1,2, YANG Xue-zhi1,2, DONG Zhang-yu1,2, SHI Cong-cong1,2 2019, 40(1): 117-123.  DOI: 10.11996/JG.j.2095-302X.2019010117 Abstract ( 134 )   PDF (1862KB) ( 215 )   Object-oriented change detection technology has been widely used in the field of high-resolution remote sensing images. As the remote sensing images are affected by imaging conditions such as illumination, atmospheric environment and other factors, the quality of image features also varies. Selecting high-quality features becomes the key of the change detection of remote sensing image at the object level. For the above problems, a change detection method of object-level remote sensing images based on Relief-PCA feature selection has been proposed. In the proposed method, first of all, the original image is multi-scaled to obtain the target object. Afterwards, the spectral features and texture features of the object are extracted. Then a logarithmic ratio method is used to obtain the change vector, and the object features of the original image are filtered and dimensioned through the Relief-PCA feature selection method. Finally, the change vector analysis (CVA) variation intensity map is calculated and generated. The Otsu method is used to conduct the threshold segmentation of the variation intensity map to obtain the final change detection result. Experimental results show that compared with other state-of-the-art methods, the proposed method has higher detection accuracy, lower misdetection rate and lower missed detection rate. Related Articles | Metrics

Select

A Graph-Based Method for Visual Analysis of Power Data LI Wen-fang1,2, CHENG Xin3，LU Qiang2,3 2019, 40(1): 124-130.  DOI: 10.11996/JG.j.2095-302X.2019010124 Abstract ( 80 )   PDF (827KB) ( 145 )   Applying visual analysis technology to the power system can effectively solve the massive power data analysis and display problems caused by the development of the power system, thereby assisting the power sector in making decisions. Based on the city power data, a graph-based power data visual analysis method is proposed. Firstly, the urban power supply data is preprocessed, then the urban power supply network is modeled. Finally, according to the visualization principle, different visualization methods are designed for the visualization and analysis of multiple power data in the urban power supply network. Based on the case study of power data in a certain area of Hefei City, it is shown that this method can effectively reflect the operating status of power systems and the distribution of power customers in a certain area. Related Articles | Metrics

Select

Research on Vanishing Point Detection of Unstructured Road  Based on Directional Texture HUANG Jun, HOU Bei-ping, DONG Fei, WU Ying-dong, LIU Yu, WANG Jian-ming 2019, 40(1): 131-136.  DOI: 10.11996/JG.j.2095-302X.2019010131 Abstract ( 96 )   PDF (2514KB) ( 144 )   With the rapid development of vehicle intelligence in these years, the automatic detection of road areas has been playing a more and more important role in the field. However, the detection of unstructured roads faces significant difficulties due to the fact that many unstructured roads do not have prominent lane marks or boundaries. Conducting the detection while setting the vanishing point as the constraint can substantially improve the performance of unstructured road detection. But in practical application, the existing methods for vanishing point detection of unstructured road generally have significant shortcomings of high computation cost and poor real-time performance. In order to improve the efficiency of calculation, a new method based on local directional pattern (LDP) texture feature for vanishing point detection of unstructured roads is proposed. Through the calculation of LDP texture features, the Kirsch mask is used to obtain the four-direction response amplitude of the pixel in picture, and the error of detection is reduced by the amplitude correction; and then the main direction of texture is obtained by calculating the corrected response amplitude; the vanishing points are selected by using the local adaptive soft voting method. Finally, the detection for vanishing points is achieved. According to the experimental results, the method is faster than existing methods and it can detect the vanishing points of unstructured roads accurately and effectively in the natural environment. Related Articles | Metrics

Select

2D-FFT Image Edge Enhancement Design Based on FPGA REN Yong-feng, WU Hao-nan, CHU Cheng-qun, JIAO Xin-quan 2019, 40(1): 137-142.  DOI: 10.11996/JG.j.2095-302X.2019010137 Abstract ( 209 )   PDF (653KB) ( 189 )   When processing the information like images, we find that image details can often convey more information, and that is why people are more concerned with images details. However, under the condition of unsatisfactory illumination, the images collected by sensors are usually with low contrast and difficult to distinguish the details. To solve the above problems, an image edge extraction and enhancement method based on field-programmable gate array (FPGA) is proposed. Four parallel 512×512-point fast Fourier transform (FFT) processors are designed by modular design, and the resource consumption in FPGA is reduced by multiplexing FFT modules. At the same time, high-pass filtering algorithm and inverse fourier transform algorithm of image spectrum are realized. The simulation and experiment show that the method is effective, reliable and real-time, and it can also meet the needs of industrial image processing. Related Articles | Metrics

Select

Bicubic Image Magnification Based on Local Interpolation JI Lin-lin1, WANG Ping2,3, ZHANG Yun-feng2,3 2019, 40(1): 143-149.  DOI: 10.11996/JG.j.2095-302X.2019010143 Abstract ( 97 )   PDF (1175KB) ( 251 )   Image details and edge features play a very important role in the visual effect of images. Therefore, a key to image zooming is to keep image details and edges. In this paper, a new method for image zooming is proposed to construct the sampling surface piecewise defined by bicubic polynomials with the constraints of distance and edge. The method consists of two steps: constructing fitting surface and modifying surface. First, in each neighborhood region where the pixel is located, the new method constructs a quadratic polynomial sampling surface fitting pixels with distance and edge constraints; the weighted combination of all the quadratic polynomial sampling surfaces produces the general surface, which is defined piecewise by the bicubic polynomials. The magnified image obtained from the general surface has higher accuracy and better visual effect. Second, a method for constructing corrected surface is presented, which can improve the quality of the enlarged image. The corrected surface is constructed by the error image that is estimated via magnified image. Moreover, in order to reduce the computational cost of the method, the new method divides the coefficients of the quadratic polynomials into three categories, and proposes an algorithm for calculating each type of coefficients. The proposed method can magnify the image at arbitrary scales. Experimental results show that the new method not only improves the peak signal to noise ratio (PSNR) and structural similarity index (SSIM), but also improves the visual effect of the image. Related Articles | Metrics

Select

Automatic Segmentation of Dragon Design Based on Bi-Level Model in Chinese Imperial Costume Images ZHAO Hai-ying1,2, YANG Ting1 2019, 40(1): 150-157.  DOI: 10.11996/JG.j.2095-302X.2019010150 Abstract ( 65 )   PDF (7599KB) ( 130 )   The design pattern of Chinese imperial costumes contains rich cultural connotation. However, due to the lack of data set of pixel-level semantic annotation, the accurate segmentation of Chinese imperial costume images has become a very challenging problem. In this paper, a bi-level model integrating deep learning and GrabCut is proposed to realize the object detection and segmentation. The characteristics of different deep convolution neural network models are analyzed, and a two-stage object detector R-FCN is selected in the object detection layer (ODL). The segmentation layer (SL) of the proposed model employs GrabCut algorithm based on graph theory to produce final segmentation result. Experiments show that the proposed bi-level model can produce good segmentation results in the Chinese imperial costume image data set. Related Articles | Metrics

Select

Multi-Objective Identification of UAV Based on Deep Residual Network ZHAI Jin-you1, DAI Ji-yang1,2, WANG Jia-qi1, YING Jin1 2019, 40(1): 158-164.  DOI: 10.11996/JG.j.2095-302X.2019010158 Abstract ( 167 )   PDF (1218KB) ( 292 )    In traditional target recognition algorithms, the classical region proposal net (RPN) has large amount of computation and high complexity of time at extracting the target candidate region. Cascade region proposal network (CRPN) is proposed as a new search method for improving the performance of RPN, in which residual learning based deep residual network (ResNet) is also used effectively to suppress the degradation phenomenon in deep-level convolution neural networks. Aimed at the network models with different depths and parameters, a novel multi-strapdown residual network (Mu-ResNet) model, which is of less memory and lower time complexity, is designed by combining two-layer and three-layer residual learning modules. The combination model of Mu-ResNet and CRPN is used for multi-target recognition test by using the unmanned aerial vehicle (UAV) target data and PASCAL VOC data. The results have shown that nearly 2% of recognition accuracy is increased compared with the combination model of ResNet and RPN. Related Articles | Metrics

Select

Approximate Soft Shadow Rendering Based on Adaptive Visibility Filtering YANG Chao-zhi, CHEN Chun-yi, QU Fu-heng, CHEN Sheng, XING Qi-wei 2019, 40(1): 165-172.  DOI: 10.11996/JG.j.2095-302X.2019010165 Abstract ( 143 )   PDF (3566KB) ( 171 )   Focusing on the difficulty in interactively rendering physically correct soft shadows existing in global illumination effects, this study proposes an algorithm for producing approximate soft shadows, which can generate alterable penumbra reflecting spatial positions of occluders. Firstly, a binary light-visibility map is generated by using occluding detection method based on ray tracing and taking the center of light source as a reference point. Secondly, a way that adaptively determines the kernel size of spatial smoothing filter is proposed. Then, visibility factors that change smoothly from visible region to non-visible region are obtained by a method of adaptive visibility filtering. The filtering method can solve the problem of visibility infiltration by using mask. Finally, the visibility factors are used to dynamically modulate the direct illumination of the associated visible points in the process of ray tracing. The highly realistic soft shadow effects are produced by adding the modulated direct illumination and ordinary indirect illumination together. Experimental results show that the algorithm is fully compatible with ray tracing and hence is easy to compute indirect illumination including mirrored visual effects. The soft shadows in images generated by our method look like physically-correct soft shadows. The rendering speed of our method is beyond 30 frames per second for all test scenes featuring specular reflections, hence satisfying the requirements of interactive applications. Related Articles | Metrics

Select

Automatic Segmentation Algorithm of Lumen in Sequence IVUS Images WANG You-sheng, WANG Yu-ting, XIA Zhang-tao, YE Hong-mei 2019, 40(1): 173-180.  DOI: 10.11996/JG.j.2095-302X.2019010173 Abstract ( 110 )   PDF (1381KB) ( 155 )   For the sequence intravascular ultrasound (IVUS) images, there are great similarities between two adjacent frames, and a new method based on sequential IVUS image registration for automatic extraction of lumen is proposed. Firstly, in order to realize the automation of extraction process, the morphological operation and the connected component method are used to extract the approximate lumen contour of the initial frame. Secondly, the target region, which includes both foreground and background pixels, is implicitly represented by a level set. The color histogram feature of target region foreground pixels and background pixels are modeled on two adjacent frames because of its merits such as independence of scaling and rotation, robustness to partial occlusions, low computational cost, etc. The Bhattacharyya coefficient measures the similarity between two adjacent frames. The higher the Bhattacharyya coefficient between target model and candidate target model is, the higher the similarity between them is. By establishing affine transformation model, the contour is more accurately located near the lumen. Finally, the segmentation procedure refines the affine transformation estimated in the registration stage, and computes the target’s true shape accurately. The target contour curve is accurately converged to the lumen of the IVUS image by the variational method and the steepest ascent method. Compared with the literature [17], the RMSE of this method is reduced by 0.124 on average, and the RDD is reduced by 0.51% on average. Compared with literature [4], the RMSE of this method is reduced by 0.063 on average, and the RDD is reduced on average by 0.16%. The experimental results show that the method can accurately extract the lumen of several consecutive frame IVUS images. Related Articles | Metrics

Select

Research on Man-Machine Interface Design Based on Scientific Experiment ——A Case Study of the Improved Design of Gree Air Conditioner Interface WANG Miao-hui1, LI Yan1, SONG Wu2, HUANG Mei-ying1, CHEN Li1 2019, 40(1): 181-185.  DOI: 10.11996/JG.j.2095-302X.2019010181 Abstract ( 176 )   PDF (826KB) ( 270 )   This study analyzes the efficiency of the man-machine interface design based on the scientific and credible industrial design theory, taking the re-design of KF_72LW operation interface as an example. A series of methods were employed in the study including questionnaire, usability test, psychological model experiment, task model experiment and other scientific methods. The framework layer of the interface is found to be unreasonable, and the distribution of the keys illogical. According to the data generated from the experment, the design problems were identified and the interface was re-designed, then the efficiency of the re-design was approved by the independent sample comparison experiment. Related Articles | Metrics

Select

Research on the Construction of Indoor Map Model Based on IFC HUANG Nan-xin1, WANG Jia1, LI Zhi2, ZHOU Xiao-ping1,3, ZHANG Chun4, SONG Bing-yu4 2019, 40(1): 186-192.  DOI: 10.11996/JG.j.2095-302X.2019010186 Abstract ( 113 )   PDF (1724KB) ( 219 )   Indoor map model construction is a basic research, which can provide data and technical support for indoor intelligent applications such as indoor navigation, emergency evacuation, and robot services. Traditional indoor information extraction methods are time-consuming and costly, and the extracted indoor information is usually incomplete. In the existing indoor map research, the model usually has a large volume, the data is complex and the redundancy is serious, and the application rate is low. With the development of BIM technology and the advancement of national policies, this study provides new ideas for indoor map model research. The paper combines BIM technology with indoor map model research, using BIM universal interactive format (industry foundation classes) IFC file as data source to extract geometric and semantic information and puts forward a new construction method of indoor map model. By defining three types of nodes in the model, the abstract expression of the internal information of the building monolayer is completed. Using the idea of node normalization, set the threshold to simplify the number of nodes in the indoor topology network and achieve the purpose of optimizing the indoor map model. In the indoor map model, the shortest path can be generated by the classic path finding algorithm, and the design algorithm will realize the path optimization. Related Articles | Metrics

Select

Dynamic Allocation of Weights to Experts for Evaluation of  Product Design and Its Application LI Xiao-ying, ZHOU Da-tao 2019, 40(1): 193-200.  DOI: 10.11996/JG.j.2095-302X.2019010193 Abstract ( 77 )   PDF (1253KB) ( 175 )   In order to evaluate product design more reliably, a process for dynamic allocation of weights to experts is proposed. The group consensus model is introduced to measure the degree of difference between expert opinions. The level of similarity is used to detect and eliminate very distinct expert opinions, thereby allowing a consensus to be reached more easily and efficiently. Weights of expert opinions are determined dynamically using the clustering algorithm to sort out the quality of various product designs. The proposed method is applied to the real-world intelligent warning sign of vehicle. Results demonstrate the ability of the proposed method to capture the difference between expert opinions effectively, allocate weights to expert opinions dynamically and evaluate the product more reliably. Finally, the proposed method is compared with other competing methods to prove its superiority, yielding insights into the evaluation of product design. Related Articles | Metrics

Select

Constructing a New System of Three-Dimensional Engineering Graphics Teaching Based on OBE ZHANG Jing-ying, YANG Wei, TONG Xian-ying, ZHANG Hui, ZHANG Tong 2019, 40(1): 201-206.  DOI: 10.11996/JG.j.2095-302X.2019010201 Abstract ( 211 )   PDF (992KB) ( 254 )   Based on a new system of integrated mechanical engineering teaching within an Engineering Graphics curriculum, following the outcome-based-education (OBE) concept as a key guideline, mainly relying on in-classroom format of teaching, integrating and exploring additional extracurricular teaching resources, this study constructs an “online-to-offline integrated teaching platform” that promotes students’ ability and motivation for learning using various “Internet+” technologies and tools. This new and all-around multi-dimensional Engineering Graphics teaching system “integrates in-class learning and after-class learning, combines online tools and offline tools, brings together mandatory courses and open-ended hands-on experiment courses, bonds teachers with students, and integrates theory and practice.” Through seven years of practical teaching experience, the system has gradually improved, and could reasonably resolve some common problems that have previously troubled lots of teachers of fundamental-level engineering curriculum--problems such as insufficient scheduled hours in class, excessively high student-teacher ratio, lack of hands-on experiments or practice, insufficient after-class supplementary assistance, and students’ relatively low self-motivation in learning. Decent teaching results have been observed through the new system, and can provide some reformatory ideas for teaching basic engineering curricula. Related Articles | Metrics

Select

Imaginary Wall Method to Determine the Visibility of Line and Plane Intersection or Plane-Plane Intersection WANG Lin, BI Yan 2019, 40(1): 207-211.  DOI: 10.11996/JG.j.2095-302X.2019010207 Abstract ( 155 )   PDF (447KB) ( 183 )   An imaginary wall method is proposed to determine the visibility of line and plane intersection or plane and plane intersection. Three typical cases are discussed concerning the former topic: a line in the general position intersects with a plane in the special position, a line in the general position intersects with a plane in the general position, and a line in the special position intersects with a plane in the general position. Three typical cases are also discussed on the latter topic: the intersection of two planes both in the special position, a plane in the general position intersects with a plane in the special position, and the intersection of two planes both in the general position. The imaginary wall method can determine the visibility of line and plane intersection or plane and plane intersection quickly and accurately. This method can also provide some references for the learning and teaching in mechanical drawing courses. Related Articles | Metrics