Transactions on Edutainment XI

As large public repositories of 3D objects continue to grow, more and more feature extraction technologies for 3D objects spring up. On the basis of classical algorithm, new factors have been added to these emerging technologies. Feature extraction technologies, which are based on shape structure and geometry information, include semantics, kinematics and cognition, etc. While the technologies have been developing, using features to solve problem is more important than just extracting features from 3D objects. In this paper, we summarize several feature extraction technologies from different aspects. Then we aim at the applications of 3D object feature, not just the general 3D models retrieval, mainly about some specific applications and target on 3D CAD objects, non-rigid 3D objects and deformable objects.

This paper proposes a method that combines Temporal Modified-RANSAC(TMR) with a fixation-based segmentation algorithm for reconstructing the 3D structure of moving and still objects that are tracked in video and depth image sequences acquired by moving Kinect© and/or range finders First, we compute 3D optical flow of feature points. Second, TMR classifies all flows into consistent 3D flow sets for the background and each moving object; simultaneously, the rotation matrix and the translation vector for each 3D flow set are computed. Then, an improved fixation-based segmentation method segments each object’s area. Finally, dense 3D models for the background and each moving object are constructed along with each object’s rotation matrix and translation vector in each frame. Experiments using multiple moving objects in color and depth image sequences acquired by Kinect(c) demonstrate the effectiveness of our proposed method.

This paper presents a method for three-dimensional reconstruction from images that can be used to identify obstacles based on a single omnidirectional image obtained by an omnidirectional stereo vision optical device (OSVOD). The OSVOD is a novel catadioptric system based on a common perspective camera coupled with two hyperbolic mirrors. The images captured by OSVOD are unwrapped into cylinder panoramic images, which are then examined for stereo matching along vertical epiploar lines. Given two images of a space point obtained by this vision system, the 3D coordinates of the point can be calculated by using triangulation. An algorithm of the quick stereo matching method based on Canny edge and SIFT feature points is proposed and the result of 3D reconstruction and depth map are obtained. From the results, we can find that the 3D reconstruction accuracy can be accepted for mobile robots’ obstacle detection and navigation tasks.

The development of image preprocessing has provided new opportunities in the field of three-dimensional reconstruction. One of the most important areas of three-dimensional reconstruction is focused on model registration by means of matching algorithm. This is mainly due to the great increase of registration algorithm in the pattern recognition system such as image acquisition, image preprocessing, 3D reconstruction. This paper presents an analysis of model registration algorithm of three-dimensional reconstruction by comparison common registration algorithm such as RANSAC (Random Sample Consensus) and ICP (Iterative Closest Point). Then, in order to elevate registration precision and robustness affecting the 3D reconstruction results, CTF (Coarse to Fine) registration strategy based on RANSAC-ICP Algorithm is proposed. Finally, by using three-dimensional reconstruction experiment based on RANSAC-ICP Algorithm, the performance of CTF registration strategy has been analyzed, and some problems and design solutions have been identified and registration precision and robustness have also been validated by experimental results.

A novel license character recognition method based on the nearest feature line (NFL) classifier is presented in this paper. Based on ocular uniformity, we process our color clustering in Munsell color space using the concepts of NBS distance. Then, with different structure elements, we take a series of operations to the image that had been processed by color clustering method before, after the location of the plate be found accurately, we propose a fast algorithm based on the NFL method to improve the performance of the recognition system by using a lot of virtual data generated by the feature line. Experiment result shows that the result of the method is efficient and gratifying.

In this paper moving image matching task allocation strategies based on the Multi-Agent system are discussed. Combined with genetic algorithm and simulated annealing algorithm, it gives out a new moving image pattern match-ing task allocation strategies based on the Multi-Agent system. Due to the unique nature of genetic algorithms and simulated annealing algorithm, makes the task allocation algorithm presented in this paper has new features that is different from the traditional method.

In this paper, we propose a new video matting method based on sampling. By detecting the movement of foreground and background objects from video, we define the local transformation which transfer the small areas between different frames in the video. By the local transformation, foreground and the occluded background object can be sampled from neighbor frames. Benefited from temporal consistency, we can get high quality alpha matte. Experiment results show that our method achieves ideal results in the scenes where old method get poor results.

Compared to free sketch, gesture-based sketch recognition can achieve high accuracy by requiring the user to learn a particular drawing style in order for shapes to be recognized. In this case, choosing an appropriate classifier is quite critical. This paper compared three different algorithms for labeling each drawn stroke as being a particular component in the generic model. Our statistic shows that K-means classifier yields better results than the other two and we test that by applying this classifier to rocket sketches.

Caricature is the art which has exaggerated feature without losing identifiable characteristics of true face. People like the caricature and the acceptance level is very high, along with the widely used cartoon. But the cartoon drawing is an artistic behavior and ordinary people has not been trained strictly, so it is very hard to draw a caricature for themselves or anyone else. Based on these, automatic generation system is born to simulate the caricature. However, the difficulty lies in feature exaggeration, which involves extracting features and establishing new exaggeration rules, and the current rules are not unified and integrated. So this paper devotes to establishing effective exaggeration rules, and puts it into practice. This paper takes the Thin Plate Spline to conduct the image distortion in order to achieve the face features exaggeration. Taking the Thin Plate Spline to distort image can achieve smooth effect, and then deformation only for exaggerating features, so it makes the computation easy. The source image after image distortion is slightly “real”, so we use Canny operator to extract face image edge, and use binary to keep texture information, which can generate sketch effect. Finally, a lot of tests demonstrate that the effect of face feature exaggeration is very good.

In order to improve the speed of texture synthesis, we propose an improved synthesis algorithm according to the Image Quilting algorithm in this paper. In this algorithm, the L-overlap region is replaced by r-overlap region to reduce the amount of the pixels used to compute the error of Sum-of-Squared Differences (SSD) and to lower the calculation redundancy. Parallel calculation is adopted to search the best synthesis patch from texture sample image and make further improvement on the synthesis speed. In addition, the area used for looking for the minimal cost path is enlarged from one patch to more patches that belong to the same row to retain or improve the synthesis effect. The experiment results show the proposed method has faster synthesis speed and lower algorithm complexity than the Image Quilting algorithm.

Plant modeling is challenging because of the beauty and complexity of plants. In this paper we develop an interactive modeling system to generate plants based on freehand sketches. The modeling process is decomposed by three parts: First the branches are created by the 2D sketches information. Then similarity branches are created by fractal. Finally a whole structure algorithm is used to produce the plants overall silhouette and density shape. The 2D freehand sketches have no deep information of branches. We use the phyllotaxis rule to obtain the deep information of branches by their rotation angles. For obtaining more natural whole silhouette, we provide a function to generate plant branch system by several similar sub-branches. The manipulation operation is simple and the simulation results are more natural and less distortion than other methods.

In constructing a parametric curve interpolating a set of data points, one of the key problems is to specify a parameter value (node) for each data point. A new method of choosing knots is presented. For each data points, the new method constructs a quadratic polynomial curve by three adjacent data points. The node parameters of the quadratic curve are determined by minimizing the second derivative of the quadratic curve. And the knot interval between two adjacent data points is determined by two quadratic curves associated with the two adjacent data points. Experiments showed that the curves constructed using the knots by the new method generally have better interpolation precision.

The early application of augmented reality (AR) was introduced. And the principle and application of mobile AR technology were discussed. Then the present application of mobile AR in e-learning was also discussed. Finally the existing problem and application prospect of mobile AR technology in e-learning were discussed. Our results indicated that the application of mobile AR technology in e-learning improves the students’ autonomous learning ability and learning interest, which is very important to personnel training.

Knowledge plays very important role in educational game. The paper is aimed to investigate a mapping mechanism for knowledge combination with game. The main achievements include: a new system of classification for educational game based on learning objectives and characters of knowledge in it is proposed, based on which, a general-element model of game and its formal representation is given; knowledge is classified into eleven categories according to the characters of game to be combined; relation between knowledge and game rules are analysis and a mapping mechanism is proposed; With the mapping from knowledge to game rules, a general knowledge model for integrating with game is constructed using senary expression which involves domain, entities, relation, sequence, theorem and skills. Finally, an application of agricultural game, Virtual Farm, is introduced to validate the mapping scheme. The experiences of game playing show the combination are effective accordingly making player learn much agricultural knowledge in an unconscious way during game playing.

Chinese teachers find it difficult to get optimal students learning outcomes for their English study. Reasons are obvious for this phenomenon in an EFL environment, where inadequate authentic language sources are provided. Students seldom have the chance to use English in daily life, thereby obstructing their motivation for study and causing anxiety when they are encountering native speakers. It’s suggested that Chinese teachers should adopt effective methods to achieve the optimal learning outcomes of the students. Using multimedia material, such as video, in the classroom is recommended in this article, coupled with the schema theory and the comprehensible input theory. Detailed strategies of how to choose the suitable video material, when to introduce the video material during the class teaching and how to design classroom activities based on the video material are given in the later discussion of this article.

Meticulous painting is one of Chinese traditional painting that has more than 2000 years history. We present an automatic technique that generates Chinese meticulous painting from image in this paper. First, we transfer the input image to grayscale, and extract the contours by using the Laplace operator. The contour image is the line-drawing draft of the meticulous painting. Second, we propose a grayscale hierarchical mixed method to produce the first layer ink-dyeing draft. The final ink-dyeing draft is produced by mixing the first layer draft and the second layer, which is created by Gaussian smooth. Then, we transfer the input image to HSV color space, and use the H, S and the grayscale of ink-dyeing draft to get the color draft. Finally, we have the meticulous painting by combining the colorful draft with rice paper texture. Experimental results demonstrate the effectiveness of our method in producing meticulous painting from image.

Xinjiang traditional folk arts and crafts design as the human history of civilization heritage, is not only a high cultural value immaterial cultural relics, but also precious material for the research of the development of the ancient human civilization. However, as the cultural heritage of the heir is gradually reducing, the protection and inheritance of intangible cultural heritage has become a new and difficult task. This paper puts forward motif gene is the basic unit of inheriting the national cultural connotation, and through the design segmentation algorithm to obtain traditional motif genes, through a variety of generation model and genetic recombination, deformation to create regenerating motif gene. To establish a new design model for protection and inheritance of traditional folk arts and crafts, and give a new model of system structure as well as the realization strategy in paper.

The promotion of the export of cultural and creative products (services) helps to adjust the trade structure of Hangzhou, drive the development of the cultural and creative industries and the transformation and update of the economic structure, enhance Hangzhou’s popularity and reputation, and improve the city image; meanwhile, it facilitates the global cultural exchange and sharing of civilization created by human beings. After the analysis of the status quo of Hangzhou’s cultural and creative industries, this paper explores the factors that influence the export of the cultural and creative industries of Hangzhou, and then discusses the problems with and measures for the export of cultural and creative products of Hangzhou by meanwhile drawing upon the experience of the countries which are advanced in the industries.

Set Pair Analysis (SPA) is a new subject which was proposed by Zhao Keqin at 1989, which is a type of theories and developed for utilizing in uncertainty system. In this paper, from the view point of operational research, we called identical discrepancy contrary analysis, identical discrepancy contrary inference and identical discrepancy contrary decision based on Set Pair Analysis together as identical discrepancy contrary mastermind, and then takes the example of Scripts and Scribes to illustrate the applications of Set Pair Analysis in identical discrepancy contrary mastermind.

In order to explore the influence of Baroque music on Undergraduates’ recognition memory, this study randomly selected 63 students from a university as subjects, taking mixed design of 3 × 2 to examine the recognition accuracy of the subjects on Chinese and English materials under the condition of Baroque music as background music as well as pilot music and without background music and collected each band EEG data of the subjects under the three conditions through biofeedback instrument. The results showed that: (1) The music group’s main effect was not significant; No matter what kind of music mode, the recognition accuracy of Chinese material was significantly higher than the English; (2) There were significant differences between groups of pilot music and background music as well as without background music in α brain waves, but there was no significant difference between group of without background music and background music.

Talents are one of keys to develop service outsourcing industry and the shortage of service outsourcing talents has been a bottleneck to restrict the development of service outsourcing industry in China. Universities should follow the national strategic planning of developing service outsourcing industry from both sides of graduate employment and cultivating service outsourcing talents. Thus, it is very important for universities to study the training of service outsourcing students. The urgency and feasibilities to training service outsourcing students through cooperation between universities and enterprises are pointed out in this paper by analyzing status of training service outsourcing talents in universities. The cooperation patterns and requirements between Hangzhou Institute of Service Engineering and enterprises in training service outsourcing students are presented.

Because of the popularity of mobile phone and the development of mobile network, mobile data is growing explosively. Mobile data mining is more and more of attention. But single node-based data mining platform has been unable to store and analysis the massive data. According to cloud computing technology, we preset a distributed data mining framework based on Hadoop. Then, we present the implementation of this system framework and process mobile internet access log on the Hadoop cluster. Comparative tests will show that this distributed system framework is significantly efficient for processing huge scale dataset.

KACA is a typical local-recoding k-anonymization algorithm. It can generate k-anonymizing data with high quality. The main drawback of KACA algorithm is its high computational cost in dealing with large dataset. To remedy this problem, we propose an new efficient k-anonymization algorithm. The main idea of the proposed algorithm is that we first adopt the c-modes algorithm to partition the whole dataset into some large clusters, and then take KACA algorithm to k-anonymize each cluster separately. Finally, comprehensive experiments demonstrate the effectiveness of our algorithm.