Time filter

Source Type

Zhang Y.-L.,Zhejiang University of Technology | Zhang Y.-L.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Zhang W.,Zhejiang University of Technology | Xiao G.,Zhejiang University of Technology | And 3 more authors.
International Conference on Control, Automation and Systems | Year: 2012

Image segmentation is an important and classic problem in image processing and computer vision. Thresholding is applied to many fields, because of its less computation and stable performance. The key of thresholding method is to determine the adaptive threshold. In order to segment biological image effectively, a new adaptive thresholding method is proposed. First, two dimension minimum entropy is computed based on gray-gradient co-occurrence matrix; and then the genetic algorithm is applied to encode the two-dimension threshold vector; Finally, the optimum threshold is calculated based on fitness function and uniformity measurement(UM). Experimental results show that this method has three advantages: 1) improve computational efficiency so that it can run in real time; 2) retain more object and edge information so that it can meet the practical requirement; 3) robust to the uneven distribution of light. © 2012 ICROS.


Liu S.,Zhejiang University of Technology | Liu S.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Jin H.,Zhejiang University of Technology | Mao X.,Zhejiang University of Technology | And 3 more authors.
The Scientific World Journal | Year: 2013

This paper proposes a segmentation-based global optimization method for depth estimation. Firstly, for obtaining accurate matching cost, the original local stereo matching approach based on self-adapting matching window is integrated with two matching cost optimization strategies aiming at handling both borders and occlusion regions. Secondly, we employ a comprehensive smooth term to satisfy diverse smoothness request in real scene. Thirdly, a selective segmentation term is used for enforcing the plane trend constraints selectively on the corresponding segments to further improve the accuracy of depth results from object level. Experiments on the Middlebury image pairs show that the proposed global optimization approach is considerably competitive with other state-of-the-art matching approaches. © 2013 Sheng Liu et al.


Tang Y.,Zhejiang University of Science and Technology | Tang Y.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Wu D.-Y.,Zhejiang University of Science and Technology | Wu D.-Y.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | And 2 more authors.
Computational and Mathematical Methods in Medicine | Year: 2013

This paper proposes a computational approach to seasonal changes of living leaves by combining the geometric deformations and textural color changes. The geometric model of a leaf is generated by triangulating the scanned image of a leaf using an optimized mesh. The triangular mesh of the leaf is deformed by the improved mass-spring model, while the deformation is controlled by setting different mass values for the vertices on the leaf model. In order to adaptively control the deformation of different regions in the leaf, the mass values of vertices are set to be in proportion to the pixels' intensities of the corresponding user-specified grayscale mask map. The geometric deformations as well as the textural color changes of a leaf are used to simulate the seasonal changing process of leaves based on Markov chain model with different environmental parameters including temperature, humidness, and time. Experimental results show that the method successfully simulates the seasonal changes of leaves. © 2013 Ying Tang et al.


Liu S.,Zhejiang University of Technology | Liu S.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Zhai B.,Zhejiang University of Technology | Chan S.,Zhejiang University of Technology | And 2 more authors.
Mathematical Problems in Engineering | Year: 2013

Based upon the framework of the structural support vector machines, this paper proposes two approaches to the depth restoration towards different scenes, that is, margin rescaling and the slack rescaling. The results show that both approaches achieve high convergence, while the slack approach yields better performance in prediction accuracy. However, due to its nondecomposability nature, the application of the slack approach is limited. This paper therefore introduces a novel approximation slack method to solve this problem, in which we propose a modified way of defining the loss functions to ensure the decomposability of the object function. During the training process, a bundle method is used to improve the computing efficiency. The results on Middlebury datasets show that proposed depth inference method solves the nondecomposability of slack scaling method and achieves relative acceptable accuracy. Our approximation approach can be an alternative for the slack scaling method to ensure efficient computation. © 2013 Sheng Liu et al.


Dong T.,Zhejiang University of Technology | Dong T.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Liu S.,Zhejiang University of Technology | Liu S.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | And 3 more authors.
PLoS ONE | Year: 2015

To automatically adapt to various hardware and software environments on different devices, this paper presents a time-critical adaptive approach for visualizing natural scenes. In this method, a simplified expression of a tree model is used for different devices. The best rendering scheme is intelligently selected to generate a particular scene by estimating the rendering time of trees based on their visual importance. Therefore, this approach can ensure the reality of natural scenes while maintaining a constant frame rate for their interactive display. To verify its effectiveness and flexibility, this method is applied in different devices, such as a desktop computer, laptop, iPad and smart phone. Applications show that the method proposed in this paper can not only adapt to devices with different computing abilities and system resources very well but can also achieve rather good visual realism and a constant frame rate for natural scenes. © 2015 Dong et al.


Fan J.,Zhejiang University of Technology | Fan J.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Shi X.Y.,Zhejiang University of Technology | Zhou Z.,Zhejiang University of Technology | Tang Y.,Zhejiang University of Technology
Science China Information Sciences | Year: 2013

The framework of texture-by-numbers (TBN) synthesizes images of global-varying patterns with intuitive user control. Previous TBN synthesis methods have difficulties in achieving high-quality synthesis results and efficiency simultaneously. This paper proposes a fast TBN synthesis method based on texture optimization, which uses global optimization to solve the controllable non-homogeneous texture synthesis problem. Our algorithm produces high quality synthesis results by combining texture optimization into TBN framework with two improvements. The initialization process is adopted to generate the initial output of the global optimization algorithm, which speeds up the algorithm's convergence rate and enhances synthesis quality. Besides distance metrics to measure image similarities are specifically designed for different images to better match human visual perception for structural patterns and a user study is conducted to verify the effectiveness of the metrics. To further improve the synthesis speed, the algorithm is entirely implemented on GPU based on CUDA architecture. The optimized TBN method is applied to various visual applications including not only traditional TBN applications, but also image in-painting and texture-based flow visualization. The experimental results show that our method synthesizes images of higher or comparable qualities with higher efficiency than other state-of-art synthesis methods. © 2013 Science China Press and Springer-Verlag Berlin Heidelberg.


Qin X.,Zhejiang University of Technology | Qin X.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Zhao H.,Zhejiang University of Technology | Zhang Q.,Zhejiang University of Technology | And 3 more authors.
Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics | Year: 2014

When processing the structured light stripes with obstructions and shadows, the conventional decoding algorithms can lead to erroneous decoding and cause problems that stripes with different coded values overlapped. In this paper an improved Canny edge-detection method to extract the boundaries between the black and white stripes is proposed, and a precise coding and decoding algorithm for structured light based on the boundaries information is presented. Considering the features of the stripes, the Canny edge-detection algorithm is improved to extract the stripes boundaries precisely. The proposed coding and decoding methods for structured light include following steps. Firstly, build codec binary tree to describe the structured light image sequence stripes boundary. Then design boundary location and decoding algorithm to solve the unknown boundary code value. Thirdly, design boundary matching algorithm to eliminate different code value stripes overlapped. Finally, decode the disconnected boundaries caused by obstructions and shadows. The experiments show that the good results can be obtained by the algorithm. It has higher decoding accuracy, fewer errors, and better robustness when processing the stripes of complex shape.


Zhu L.-N.,Zhejiang University of Technology | Zhu L.-N.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Zhao Y.-W.,Zhejiang University of Technology | Wang W.-L.,Zhejiang University of Technology | Wang W.-L.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province
Jisuanji Jicheng Zhizao Xitong/Computer Integrated Manufacturing Systems, CIMS | Year: 2012

To issue and update cloud information in real time and to find and obtain resource information rapidly under massive data environment of cloud manufacturing, three-attribute structure for cloud manufacturing was built based on dynamic attributes of resources and services, and Resource Via Cloud Service(RVCS) based resource package, publication and discovery model was proposed. The realization process of dynamic attribute publication, Web Service Description Language(WSDL), document structure, resource cloud publication and its discovery as well as matching was introduced in detail. The prototype system was constructed by using mpiBLAST and Microsoft .Net. Thus, the basic function was realized. It provided theoretical and technical supports for further research.


Qin X.,Zhejiang University of Technology | Qin X.,Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province | Wang H.,Zhejiang University of Technology | Du Y.,Zhejiang University of Technology | And 3 more authors.
Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics | Year: 2013

In structured light geometry reconstruction, as the projecting modes and lighting conditions are complex and changeful, detail information in dark areas are usually lost in the result images. An improved Retinex algorithm are proposed based on HSV color space with color restoration and color saturation correction strategy. According to the requirement of color retain, the algorithm includes following steps. Firstly, an input color image is converted from RGB color space into HSV color space. Then the traditional multi-scale Retinex algorithm is applied only to V-component through the analysis of HSV color space model. At the same time, the coefficient of correlation is used to adaptively adjust the S-component base in the enhancement of V-component. Finally by transforming HSI model into RGB model, the enhanced color image with color restoration is obtained. Experimental results show that, the algorithm used in structured light stripe image enhancement, the color of the structured light image is maintained and the detail information is enhanced very well, which is more favorable to follow-up stripe information extraction and automatic coding.


PubMed | Zhejiang University of Technology and Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province
Type: Journal Article | Journal: PloS one | Year: 2015

To automatically adapt to various hardware and software environments on different devices, this paper presents a time-critical adaptive approach for visualizing natural scenes. In this method, a simplified expression of a tree model is used for different devices. The best rendering scheme is intelligently selected to generate a particular scene by estimating the rendering time of trees based on their visual importance. Therefore, this approach can ensure the reality of natural scenes while maintaining a constant frame rate for their interactive display. To verify its effectiveness and flexibility, this method is applied in different devices, such as a desktop computer, laptop, iPad and smart phone. Applications show that the method proposed in this paper can not only adapt to devices with different computing abilities and system resources very well but can also achieve rather good visual realism and a constant frame rate for natural scenes.

Loading Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province collaborators
Loading Key Laboratory of Visual Media Intelligent Processing Technology of Zhejiang Province collaborators