Search

Skull registration is one of the important steps in craniofacial reconstruction, and its registration accuracy and efficiency have an important impact on the reconstruction results. To solve the problem of low accuracy and efficiency of existing skull registration methods, a hierarchical skull point cloud registration method is proposed in this paper. The whole registration process is divided into a rough registration stage and a fine registration stage. Firstly, feature points are extracted from the pre-processed skull point cloud model, and a local coordinate reference system is established according to the feature points and their neighbor points. The improved spin image is used to construct the local feature descriptor. The feature matching is carried out according to the nearest neighbor algorithm, and the k-means algorithm is used to eliminate the mismatching points to achieve skull rough registration. Then, based on rough registration, we use an improved ICP algorithm to achieve fine registration of the skull. In this process, we use random sampling to reduce the search scale of points and add geometric feature constraints to further eliminate mismatched points. Finally, the whole registration algorithm is applied to the skull point cloud data to verify. The experimental results show that, compared with other methods, the registration effect and efficiency of the proposed method are superior to those of other methods. In order to verify the universality of the method, we also use a common data set for verification. Experiments show that the method is also very effective.

The precise segmentation of cerebral vessels is essential for the detection of cerebrovascular diseases. The complex structures of cerebral vessels and the low contranst of thin vessels in medical images make precise segmentation difficult. In this study, we propose a new phase-field and statistical model for blood vessel segmentation. The proposed model is based on the Allen-Chan equation with double well potential and statistical distribution function. The brain tissues in the image are modeled by Gaussian distribution while cerebral vessels are modeled by uniform distribution respectively. The region distribution information combined with the phase-field model is used in curve evolution. And the level set method is developed to implement the curve evolution to assure high efficiency of the cerebrovascular segmentation. Comparisons with the LBF model and LCV model show that our model can achieve better results with fewer iteration number and less time.

Due to increasing attraction of motion capture systems technology and the usage of captured data in wide range of research-oriented applications, a framework has developed as an improved version of MOCAP TOOLBOX in Matlab platform. Firstly, we have introduced a faithful script to deal with public motion capture data, which will be friendly for us. Various functions through dynamic programming, by using the Body Segment Parameters (BSP) are edited and they configured the position of markers according to data. It is used to visualize and refine without the MLS view and the C3D editor software. It has opened a valuable way of sensor data in many research aspects as gait movements, marker analysis, compression and motion pattern, bioinformatics, and animation. As a result, performed on CMU and ACCAD public mocap data, and achieved higher corrected configuration scheme of 3D markers when compared with the prior art, especially for C3D file. Another distinction of this work is that it handles the extra markers distortion, and provides the meaningful way to use captured data.

Color attention algorithm can improve the accuracy of object recognition significantly. In the algorithm, there are 3 parameters related to bin number of color features and shape features and the weight between color and shape, need to be set. Whether the parameters are set correctly or not poses a great impact on the accuracy of object recognition. However, searching suitable values in the parameters space related to color attention algorithm is a NP-hard problem and the parameters of the algorithm are set manually in recent research through which the right values are not guaranteed to be found. Recently, bio-inspired computing gains much attention for its advantages in complex optimization problems and differential evolution outperforms most other bio-inspired computing algorithms in divergence and stability in optimization problems. When taking the accuracy of object recognition as a fitness function, increasing the accuracy of object recognition is then an optimization problem to find the largest accuracy in parameter spaces. Therefore, we design the structure of the agent of differential evolution and take the classification accuracy with support vector machine algorithm as a fitness function. Then we use differential evolution to search the parameters space and find some suitable parameters for color attention algorithm successfully. Our experimental evaluation demonstrates that the accuracy of object recognition increases greatly with the right parameters.

Way of walking is a spatio-temporal phenomenon that typifies the motion characteristics of human. In this paper, we propose a human gender identification method based on the outdoor surveillance of human gait using statistical techniques and geometrical function applied to three-dimensional (3D) joint movement data. The statistical techniques are used to define the features of the joint of the human gait, and the geometrical function applied for gender recognition. Our proposed scheme is based on extraction of the concern joint (hip joint) data that provides plentiful information for gender recognition. Here the rotation angle data of a hip joint were computed from the Biovision Hierarchical data (BVH file). The use of BVH file for human gender recognition is a novel feature of our work. The results indicated that the proposed approach is highly reliable for gender recognition. DOI: http://dx.doi.org/10.11591/telkomnika.v11i6.2274 Full Text: PDF

This paper proposed a cerebrovascular segmentation model based on region growing and level set. Firstly, a statistics and region growing model is used to detect the vessel region. During the process, a Maximum Intensity Projection (MIP) image from volume data is segmented by Otsu algorithm to get the seeds, and then to get the contour of vessel by an improved region growing algorithm. Moreover, an improved local-adaptive level set method is developed to implement the accurate segmentation. It can be seen from the results that this hybrid segmentation algorithm is more accurate than the general level set algorithm, especially segmented the vessels with small radius.

For the 3D brain MRI parameterized statistical model segmentation method, the pre-processing method of brain image is brought forward for the model. First the DWM(directional weighted median) filter and Gaussian template are used to de-noise the brain MRI image. Then the Laplacian operator is used to sharpen the image, and the Robert operator is used to realize the edge detection, which can improve the measurement accuracy. And the MIP (maximum intensity projection) algorithm is applied to extract the largest connected component. The image of only brain vessel and part brain tissue can be obtained and the influence of background and no brain vessels like tissue to the images is eliminate. After pre-processing stage, the brain image is analyzed as the input to the parameterized statistical model. The small branches of the brain vessel can be segmented.

Because the greatly most algorithms of real color image enhancement did not relate to the spectral sensitivity of most human vision, the color restoration functions of some real color image enhancement algorithms are greatly at random and not proved , and the real color images enhanced which are based on illumination-reflectance model have the loss of details and the ‘halos’, we proposed a new algorithm to overcome these disadvantages. Firstly, we transform the real color image from RGB space to HSV space which is approximately orthonormal system between color and value. Secondly, the illumination and the reflectance of value are separated by homomorphic decomposition based on illumination-reflectance model. We have discovered that the high dynamic range of image including high bright lights is mainly caused by the reflectance. Thirdly, the details of reflectance are preserved by wavelet transform. Fourthly, the dynamic range of reflectance is compressed by Butterworth filtering. Lastly, the energy of the saturation of real color image in HSV space is attenuated according to the spectral sensitivity of most human vision. Experiments show that the color of image enhanced adapts well to the most people vision and is not obviously distorted, the preservation of the details is perfectly and the ‘halos’ is effectively restrained. Real color images enhanced by the algorithm are much better than the multi-scales Retinex’s with color restoration (MSRCR).

In this paper, an effective acoustic signal processing method has been presented in research of Musical Information Retrieval (MIR). Firstly, the raw acoustic signal was pretreated by a adaptive wavelet soft-threshold of gain factor and teager energy operator (AWSTD_GF&TEO), then it needs to extract the fundamental frequency form it. Secondly, proposed a new effective "cabinet" method to filter the waveform, after that, the waveform is available compared with the one unfiltered, at last, implemented the retrieval operations with an improved melody representation method and string fuzzy matching method. The evaluation result shows that the AWSTD_GF&TEO and the new "cabinet" method could help to retrieval the object more accurate.

Because the color restoration functions of real color image enhancement are greatly at random and not proved, and the enhanced real color images which are based on illumination-reflectance model have the loss of details and the 'halos', we proposed a new algorithm to overcome these disadvantages. Firstly, we transform the real color image from RGB space to HSV space which is approximately orthonormal between color and value, and keep the color of image. Secondly, the illumination and the reflectance of value are separated by homomorphic decomposition based on illumination-reflectance model. We have discovered that the high dynamic range of image including high bright lights is mainly caused by the reflectance. Thirdly, the details of reflectance are preserved by wavelet transform. Lastly, the dynamic range of reflectance is compressed by Butterworth filtering. Experiments show that the color of image enhanced is not obviously distorted, the preservation of the details is good and the 'halos' is effectively restrained. Real color image enhanced by the algorithm is much better than the multi-scale Retinex's with color restoration (MSRCR).

In this paper we describe an ontology based framework for describing VR Objects and VR environment, reasoning and building of virtual learning environment scenario. A method is firstly proposed to semantically annotate according to the learning domain knowledge. Four kinds of semantic models are presented to describe the learning environment from different viewpoints.A simple semantic script language is suggested in order that we can describe a learning environment easily. The mapping method of the semantic modal and the VRML modal is advised. We also discuss the design and implementation technology based on our approach.

Because the most algorithms of real color image enhancement did not relate to the spectral sensitivity of most people vision, the color restoration functions of some real color image enhancement algorithms are greatly at random and not proved , and the enhanced real color images which are based on illumination-reflectance model have the loss of details and the ‘halos’, we proposed a new algorithm to overcome these disadvantages. Firstly, we transform the real color image from RGB space to HSV space which is approximately orthonormal between the color and the value. Secondly, the illumination and the reflectance of the value are separated based on a new illumination-reflectance model described by stationary wavelet transform. We have obtained the conclusion that the illumination of image is mainly preserved in the low frequency part of stationary wavelet decomposition and the reflectance is mainly preserved in the high frequency parts. Thirdly, the dynamic range of illumination is compressed by Gaussian filtering. Lastly, the energy of the saturation of real color image in HSV space is attenuated according to the spectral sensitivity of most people vision. Experiments show that the color of image enhanced adapts well to the most people vision and is not obviously distorted, the preservation of the details is excellent and the ‘halos’ is effectively restrained. Real color image enhanced by the algorithm is much better than by the multi-scales Retinex's with color restoration (MSRCR).

Digital image enhancement provides good dynamic range compression in each red, green and blue (RGB)channel, but it fails in color rendition. In this letter, we propose to transform a real-color image from RGB space to HSV space at first.Then, saturation channel is enhanced by Butterworth high-pass homomorphic filter and value channel is enhanced by multi-scale Gaussian high-pass homomorphic filters. In addition , we propose anew method to fix on the cut-off frequency of high-pass filter in homomorphic enhancement algorithm and explain the ratio coefficient of high-pass filter.Experiments show the compression of real-color image dynamic range is obviously effective, and it successes in color rendition.

The goal of adding content-based 3d model retrieval to semantic web is to develop a rich set of 3d model semantic representation to enable both humans and machines to generate and understand model descriptions and processing which can be used to enable fast efficient retrieval from model collections. The purpose of Ontology for Content-based 3d Model retrieval is intended to describe model information regardless of storage, feature extraction, creation. In our research, the ontology includes the information on media features, low level visual descriptors and their relationship and non media features of 3d model. It is represented in RDF Schema and implemented in protege.

A novel watermark algorithm is proposed for copyright protection of JPEG2000 compressed cultural heritage images in digital museum system. Visually meaningful color logo is encoded as the watermark information. Due to its characteristics, the key problem is to decide how to hide as many bits information as possible into the image while ensuring the information can be retrieved correctly, even after various attacks. In this paper, a new method is proposed to resolve it. For the frequency domain of color logo after Daubechies9/7 wavelet transform, just only low frequency and intermediate frequency coefficient is chosen as the watermark code. In order to ensure both the robust and transparency, intermediate frequency coefficient after Daubechies9/7 wavelet transform is also proposed to embedding watermark code. To achieve the lowest computational cost and provide security, the lifting scheme was presented to implement the wavelet transform. The central contribution of this paper is to develop a watermark algorithm, resilient to like JPEG2000 lossy compression attack and other mainstream attacks, embedding compressed color logo image as watermark and extracting it correctly. Experimental results show that the proposed watermark scheme is computational and quite robust in face of various attacks such as JPEG2000 compression with high compression ratios, filtering, cropping, etc.

The automatic matching method of two curves is the central important issue in the fields of computer vision, automation and pattern recognition. The method can be widely used in the object dynamic track, content-based image retrieval and the automatic assembly of robot. In order to solve the easy serious error and lacking global description of the traditional method, the periodic curvature function is brought forward in the paper. The function is raised based on the concept of continues curvature. The curvature of the continuous function not only has the unchanged character among the circumvolving, translation and zooming but also has the global character of the curves. At the same time the Gaussian function is used to smooth the curves to improve the matching precision. In the experiment, the feasibility of the method is verified further. The method has the advantages of easy to realize and quick speed, and can be easily expand to the other similar systems.