Xiaoqing Yu

Shanghai University, Shanghai, Shanghai Shi, China

Are you Xiaoqing Yu?

Claim your profile

Publications (71)1.51 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we present a robust reconstruction framework on noisy and large point cloud data. Though Poisson reconstruction performs well in recovering the surface from noisy point cloud data, it's problematic to reconstruct underlying surface from large cloud data, especially on a general processor. An inaccurate estimation of point normal for noisy and large dataset would result in local distortion on the reconstructed mesh. We adopt a systematical combination of Poisson-disk sampling, normal estimation and Poisson reconstruction to avoid the inaccuracy of normal calculated from k-nearest neighbors. With the fewer dataset obtained by sampling on original points, the normal estimated is more reliable for subsequent Poisson reconstruction and the time spent in normal estimation and reconstruction is much less. We demonstrate the effectiveness of the framework in recovering topology and geometry information when dealing with point cloud data from real world. The experiment results indicate that the framework is superior to Poisson reconstruction directly on raw point dataset in the aspects of time consumption and visual fidelity.
    2014 14th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid); 05/2014
  • [Show abstract] [Hide abstract]
    ABSTRACT: 3D visualization and real-time rendering of large-scale scenes is an important task of virtual reality. In this paper, different from local loading and roaming of the scenes, we focus on dynamic loading and real-time roaming, based on the remote servers and clients, to facilitate the efficient transmission and reduce the transmission time. More specifically, a novel dynamic scheduling algorithm in the client side is used to optimize the loading and real-time rendering performance, i.e., this method can dynamically load and unload the partitioned data blocks from the server side according to the roaming viewpoints so that we can realize the infinite roaming of a large scale scene. In order to accommodate different networking scenarios, we also design a multi-resolution scheme for those server-stored scene blocks according to different communication channel conditions, so that the most suitable resolution scheme can be automatically chosen. As shown in the experiments, our method only requires a small amount of memory overhead, while efficiently realizes infinite roaming of the large city scenes.
    Journal of Signal Processing Systems 04/2014; 75(1):15-21. · 0.55 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the rapid development of digital media technology, audio resource has become one important sources of information, but people may be confused because they do not know how to handle these resources. In this paper, we proposed a new audio fingerprinting system based on Shazam's algorithm to help people manage and find out the wanted audio resources. Salient points, not only the maxima but also the points have the largest increases in energy, have been extracted in the chromagram to construct the audio fingerprints. The idea which is similar to Constant Q Transform is adopted to optimize the salient points before the fingerprinting construction. The experiment results show that our system can recognize the wanted audio clips at the accuracy about 90%.
    IET International Conference on Smart and Sustainable City 2013 (ICSSC 2013); 01/2013
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposes a novel algorithm for high-dimensional unsupervised reduction from intrinsic Bayesian model. The proposed algorithm is to assume that the pixel reflectance results from nonlinear combinations of pure component spectra contaminated by additive noise. The constraints are naturally expressed in intrinsic Bayesian literature by using appropriate abundance prior distributions. The posterior distributions of the unknown model parameters are then derived. The proposed algorithm consists of intrinsic Bayesian inductive cognition part and hierarchical reduction algorithm model part. The algorithm has several advantages over traditional distance based on Bayesian reduction algorithms. The proposed reduction algorithm from intrinsic Bayesian inductive cognitive model is used to decide which dimensions are advantageous and to output the recommended dimensions of the hyperspectral image. The algorithm can be interpreted as a novel fast reduction inference method for intrinsic Bayesian inductive cognitive model. We describe procedures for learning the model hyperparameters, computing the dimensions distribution, and extensions to the intrinsic Bayesian inductive cognition model. Experimental results on hyperspectral data demonstrate robust and useful properties of the proposed reduction algorithm.
    Neurocomputing. 12/2012; 98:143–150.
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper proposed an improved objective audio quality estimation system compatible with PEAQ (Perceptual Evaluation of Audio Quality). Based on the computational auditory model, we used a novel psychoacoustic model to assess the quality of highly impaired audio. We also applied the robust linear MOA (Least-squares Weight Vector algorithm) and MinmaxMOA (Minmax-Optimized MOV Selection algorithm) to cognitive model of the estimation system. Compared to the PEAQ advanced version, the proposed estimation system has a considerable improvement in performance both in terms of the correlation and MSE (Mean Square Error). By combining the computational auditory model and PEAQ, our estimation system can be applied to the quality assessment of highly impaired audio.
    Proceedings of the 19th international conference on Neural Information Processing - Volume Part IV; 11/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: To dynamically load the large scaled 3D city model based on OSG (OpenSceneGraph), the model is firstly separated small blocks so as to improve the loading speed and save the memory, here we directly separate the scene along the x-axis and y-axis, then we design the algorithm to control the numbers and the positions of loading blocks, thirdly we optimize the algorithm. In the end we give the experimental result to show the virtues of our methods.
    2012 International Conference on Audio, Language and Image Processing (ICALIP); 07/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: Audio fingerprint is a content-based compact feature that summarizes an audio clip. As same as we using human fingerprint to recognize people, audio fingerprint can be used to retrieval unknown audio clips. In this paper, we propose an algorithm which extracts audio fingerprint based on dynamic subband locating and normalized SSC (Spectral Subband Centroid). SSC has been found to be related the human sensation of the brightness of an audio signal. By using this algorithm, recognize an unknown audio clip efficient and highly robust in the presence of noise be possible. The preliminary result of our experiments shows prominent performance.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: A theoretical framework of compressed sensing are represented, for real, no-spare especial audio signals, the short-term stability of these signals is analyzed, these signals are compressed using compressed sensing method, the performance of different reconstruction algorithms are studied for original audio signals based on the number of measurements and the length of frame. The experiment results show that the reconstruction performance associated with the number of measurements and the length of frames. The best reconstruction performance in all algorithms we used is OMP, the speed calculation of OMP also is good.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we realize the roaming of large scale city scene based on OSG. For large scale city, we separate it to blocks and put the blocks to the server, in order to realize rendering in client, we need to transmit the needed block to client and then render them. But the loading from server to client is a time-consuming course. General ways can't guarantee fluent roaming, here we adopt dual threads to fulfill this attitude, one thread is responsible for data transmission through net, and the other is responsible for the block's rendering. Through the coordinating work of the two threads, we realize the fluent rendering of the large city scene.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • Jie Hu, Wang gen Wan, Xiaoqing Yu
    [Show abstract] [Hide abstract]
    ABSTRACT: Pathfinding in the context of video games concerns the way in which a moving entity finds a path around obstacles; the most frequent context is real-time strategy games (in which the player directs units around a play area containing obstacles), but forms of this are found in most modern video games. It is probably the most popular but frustrating game artificial intelligence (AI) problem in game industry. This article introduces A-Star algorithm first, then describes the details of A-Star algorithm. Finally, a real example of how the A-Star pathfinding technique is used in a famous 3D game engine: Unity3D.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: The progressive mesh compression algorithm on the blocks of model based on edge-collapse mesh simplification is proposed to solve some practical problems. At first, we adopt K-means to cluster the mesh triangles. Secondly, we compress each model block. The detail information of different groups obtained by compressing each of blocks are independent of each other. If some details are lost in transmission, other details can also decode independently. We only need to request server to resend the lost part of the groups data. At the same time, it can be decoded regional details depend on our viewpoint when we need to load the complex models. Because our vision can only focus on part region of model, and the area in which we isn't interested can be instead by coarse model.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we provide a new method for rendering of large-scale 3D terrain point cloud based on bounding volume hierarchy (BVH), level of detail (LOD) and out-of-core algorithm. First, large-scale 3D terrain point clouds are partitioned into several same size small blocks according to its size. Then small blocks are reconstructed to mesh models and added texture mappings. Thirdly, we use BVH to organize scene data and create two levels to handle node loading and computer graphics rendering separately. Fourthly, we use view-dependent out-of-core method to determine which part of meshes should be rendering or deleted from the memory. In the last of the paper, we use terrain models reconstructed from point clouds to test that our method has higher usage efficiency of computer memory and rendering rates.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: With cities' sharp development, urban safety, disaster prevention, all kinds of emergency plans and post-measures have been taken great attention. This paper totally based on ArcGIS, focuses on the analysis of polluted building surfaces after threat substances exploded around urban area. The structure is as follows. To begin with, the real 3D virtual city area is rapidly created based on ArcGIS. Next, at explosion site, realistic explosion point and threat area are simulated in view. And then, by computing and marking the intersection between buildings and the threat area indeed and clearly, we can detect out the polluted building surfaces.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the development of society, the scale of modern cities has been larger and larger and the structure also been more complicated, which truly bring trouble to urban planning and management. However, it will offer great help to city decision makers by making 3D effect of key areas vivid reappearance on screen. This paper presents a method to rebuild and display Lujiazui area of Shanghai fast based on ArcGlobe, from which we can get accurate 3D image online as base map. By storing and calling models that are gathered from Goole 3D Warehouse and edited in SketchUp through symbol system in ArcGlobe, a vivid Lujiazui area will come into being. This will lay a foundation for further research of 3D virtual city.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • Jie Hu, Wanggen Wan, Rui Wang, Xiaoqing Yu
    [Show abstract] [Hide abstract]
    ABSTRACT: There exist many problems in the development of cities, such as: transportation, energy, health care, public safety, etc. It is worth for us to make some valuable researches to solve these issues. Fortunately, with the development of communication technology, microelectronics, graphics rendering technology, we can use these technologies to build a smart, efficient city for the city's future development plans and timely, effectively response to emergencies or disasters. Regardless from a technical or a political point of view, we need to make our city smarter, prosperity and progress. In this paper, we build a virtual reality system with sensor network. Make an attempt to use OpenSceneGraph, a 3D visualization platform to render data which is received by the perception of the external sensor network through network communication in real time to realize the interaction between virtual and reality world. Combining sensor network, virtual reality and network communication will create a suitable multi-dimensional information space and have a wide applicative prospect.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: Audio fingerprinting, like human fingerprint, identifies audio clips from a large number of databases successfully, even when the audio signals are slightly or seriously distorted. In the paper, based on 2-D, we propose an improved audio fingerprinting extraction algorithm which was proposed by Shazam company. The algorithm uses a combinatorial hashed time-frequency analysis of the audio, yielding unusual properties in which multiple tracks mixed together may each be identified. The results of experiment verify the improvement in the retrieval speed and accuracy.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the rapid expansion of modern multimedia data, a number of audio fingerprinting algorithms have been proposed. Audio fingerprint is a compact unique content-based digital signature of an audio signal, which can be used to identify unknown audio clips. Due to the interference of different kinds of noise, audio fingerprinting is still a challenging task. In this paper, Nearest Neighbor Estimation (NNE) is used to reduce the interference of the noise. Firstly, audio feature points are extracted from audio clips. Then NNE is used to reduce the impact of noise on the feature points. Experimental results show that NNE reduces the influence of noise effectively in white noise environment.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the rapid expansion of modern multimedia data, accurate and efficient retrieval of audio information has become one of the most important issues that need to be resolved especially in the actual noise environment. As an important part of the multimedia information, audio data provides an indispensable element for the people of auditory perception. However, in the transmission process, it will inevitably be subject to a variety of noise interference. There are generally two kinds of noise, fully distributed type and burst mixed type. In this paper, we mainly focused on noise removal of the latter kind. Combined with commonly used Philips fingerprinting algorithm, noise-removed clean audio clips are used to extract fingerprints and put into the process of retrieval. The results show that our methods can better locate noise, and get a relatively high similarity and matching accuracy.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the promotion and quick development of the concept of smart city, based on this concept how to predict and eliminate the nature disaster has been more and more possible. This paper gives a outline of how to use OpenGL to realize a system of flood simulation. First using OSGDEM to make a real terrain scene with texture and coordination. Second, based on OpenGL graphic library to simulate the water and the movement. Third, using MFC to realize the visualization of the flood.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: In audio fingerprinting, an audio clip must be recognized by matching an extracted fingerprint to a database of previously computed fingerprints. The fingerprints should reduce the dimensionality of the input significantly, provide discrimination among different audio clips, and, at the same time, be invariant to distorted versions of the same audio clip. In this paper, we design fingerprints addressing the above issues by extracting the audio fingerprints from the Spectral Flux of the clipped signal. Spectral Flux (SF) is a measure of how quickly the power spectrum of a signal is changing, calculated by comparing the power spectrum for one frame against the power spectrum from the previous frame. More precisely, it is usually calculated as the 2-norm (also known as the Euclidean distance) between the two normalised spectra. By using the AF as the feature of our algorithm we retrieval the audio clips from the database which has store some fingerprints computed previously. We test the robustness of the fingerprints under a large number of distortions. And the experimental results show that the proposed algorithm performance well in audio retrieval.
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on; 01/2012