最新刊期
卷 10 ， 期 10 ， 2005
- 摘要：Semantics-based image retrieval research comes forth for filling the gap between images' low level vision features and users' high level semantics,and image's semantic description and extraction are its crucial problems.An approach of GIS semantics-based remote sensing image retrieval is proposed.This approach mainly includes two parts: the description of spatial features' semantics and semantic matching.Object-oriented GIS semantic model and conceptual semantic network are simultaneously applied to describe the spatial features' semantics.A semantic mediator is designed to process the semantic contradiction between user and system.The system will execute a group of atomic queries to get relevant results.Assembled through Boolean calculation,these results will afterward lead to vector GIS retrieval results.Based on vector query results,by reading the remote sensing image data which have the same coordinate frame with GIS data from areas determined by the minimum exterior rectangle obtained as the union of all vector polygons,final GISSBIR retrieval results shall be produced.This approach is applied to retrieve high resolution remote sensing image database,and the results are satisfied.关键词：GIS semantics-based remote sensing image retrieval;concept semantic network;semantic mediator;object oriented GIS semantic model2755|206|0更新时间：2024-05-07
- 摘要：One of features in JPEG2000 is ROI(region of interesting) coding technique.Since the shape of interesting region is manually optional,the influence of uninterested region could be very small.This paper presents a novel and efficient scheme for remote sensing image retrieval,which does not need to decode JPEG2000's code stream completely.It extracts the spectral features of objects based on the properties that objects could reflect different waves in different wave band.The subsequent retrieval is based on the spectral feature.Further more,a new measurement scheme designed by which similarity between two images is computed and then the retrieval is realized.The experimental results show that this method is accurate and efficient.In addition,it costs much less time than the traditional method.关键词：JPEG2000;remote sensing image;retrieval of ROI2420|177|0更新时间：2024-05-07
- 摘要：A fast and efficient algorithm of automatic extraction of roads from low resolution SAR images is described here.First,a road feature detector is applied to the images to obtain the edges of roads.Then a series of templates is used for labeling the edge pixels and linking the short lines.Finally by the dynamic programming technique,the gaps between the curves of candidate roads are connected.Low-resolution SAR images of RadarSat are used to illustrate our method,and the performance is satisfactory.关键词：synthetic aperture radar image;road extraction;edge detection;dynamic programming3031|220|0更新时间：2024-05-07
- 摘要：The city-building merging algorithm is important in automated generalization of large-scale map.After discussing the shortcomings of current algorithms,a new method based on the technique of agent,TIN and clustering is put forward,which is expected to improve the operation speed,intelligentized and roboticized level of generalization.First,a clustering analysis with map data is done in order to divide whole data into several parts.Second,this algorithm regards Delaunay triangle as its basic technology,and takes a new classifying criterion to it.Third,based on the TIN classifying,an agent layer is expressed.And for every different agent layers,different agent life cycles and data structures are defined.Finally,as the final goal,a new algorithm is evolved from the techniques of agent,TIN technique of graphics calculation,and the map generalization constrains.An example and relevant analysis are presented in the end.The result shows that this algorithm is more excellent than others not only in functions but also in the speed,intelligence and automatization level.On the other hand,some modification to the algorithm is discussed.关键词：map generalization;clustering analysis;merge;Delaunay triangulation;Agent2927|238|0更新时间：2024-05-07
- 摘要：In repeat-pass synthetic aperture radar interferometry,the decorrelation caused by azimuth and range spectral shifts leads to a deterioration of the quality of SAR interferogram.There is a need to put proper spectral filters on the main and slave images.The paper discusses a new spectral filtering procedure that performs filtering before and after fine coregistration according to spectral characteristics.The procedure will reduce spectral misalignment,improve the accuracy of fine coregistration,and at last obtain high quality of interfeorgram.Four pairs of ERS-1/2 SLC data with different time interval and nominal baseline are used to analyze the validity of our method.Results show that the average values of coherence after spectral filtering flow are higher than that before spectral filtering flow,moreover the amount of residues is highly decreased.These give evidence to the goodness of the procedure.关键词：azimuth;range;spectral filtering2827|190|0更新时间：2024-05-07
- 摘要：With considerable development of the network and multimedia technology,the security problem of the video data becomes more urgent.In this paper,the scrambling strategy and methods to hide the video data are discussed deeply in the spatial and frequency domains,followed by the security of these methods and the principles of the scrambling.The means to increase the security of the scrambling methods and reduce the influence of compression efficiency and video quality due to the scrambling are analyzed in details.Based on these,the existing methods are improved in the experiments.The experimental results show that the combination of spatial and frequency domain methods enhances the security of video data.Differing from the traditional encryption methods,these scrambling methods being proposed can be embedded into the video encoding and decoding processes compactly.关键词：video data hiding;MPEG- 4;scrambling3063|179|0更新时间：2024-05-07
- 摘要：In this paper a new image interpolation algorithm is proposed.Bicubic spline interpolation method is used in the algorithm to modify the image scale.The algorithm also considers tangent vector of the edge and position revision of the interpolation point.Thus the edge propert is maintained.The experiment results show that scaled image processed with proposed algrithm is of good quality.关键词：image interpolate;bicubic spline function;edge property2849|201|0更新时间：2024-05-07
- 摘要：This paper addresses the problem of noise reduction under non-stationary environments based on the wavelet transform.This algorithm can overcome the deficiency of the conventional algorithms of noise suppression,which were only efficient for stationary environments and had large level of residual noise.The algorithm addressed in this paper is based on the different amplitude value change of image and noise and their distributing character in the wavelet domain,by this way,we can find the site and the value of the noise in the wavelet domain by making use of the iterativeness algorithm.Further,noise suppression of the image is implemented.Experiments confirm that the PSNR after denoising with the proposed algorithm is larger than the conventional algorithm;moreover,the high-frequency information of the image contains much information after the noise suppression.At the same time,the noise reduction by proposed algorithm is effective to reduce the noise under both stationary and non-stationary environments.关键词：image;noise reduction;wavelet transform;non-stationary;Lipschitz exponent2268|186|0更新时间：2024-05-07
- 摘要：Transferring color between images is a method that borrows one image's color characteristics from another.This paper introduces a technique for transferring color with auto selecting swatches.It uses unsupervised clustering to partition the target and the source image data into a certain number of subsets with similar color,then selects certain larger density pixels in each subset to construct corresponding swatch.It creates optimized correspondence by matching texture character and luminance between subsets.The color information is transferred from the matched pixel in source swatch to the target one.When color transmission is done between the source and the target swatches,the final colors are assigned to each pixel of the target image by matching pixel to pixel in the target swatches in the way of comparing texture characteristics.The proposed auto selecting swatches algorithm can be used to transfer color not only between two color images but also between a grey image and a color image.Because this technique can automatically perform color transfer process without user selecting swatches,it can be conveniently applied to batch process and video,provided that composition,texture and luminance are sufficiently distinct.关键词：auto selecting swatches;color transfer;swatches;clustering3318|245|0更新时间：2024-05-07
- 摘要：This paper presents a clustering segmentation approach for color image based on fuzzy entropy and RPCL.It not only can adaptively detect the appropriate number and centers of the initial clusters of color image for RPCL and improve the learning rate,but also avoid over-segmentation caused by fuzzy entropy thresholding approach.Firstly fuzzy entropy of each color component is computed and initial clusters' centers of each color component are determined according to the fuzzy entropy curve.Then,these centers of different color components are combined to form the initial clusters' centers of color image.But the number of these combined clusters may be larger than that of the actual clusters,which may result in the over-segmentation.Therefore,RPCL is utilized to converge some of initial centers to actual centers of original color image and image is segmented by these learned cluster centers.The experiment shows that the method can effectively and adaptively segment color images without specifying the number and centers of initial clusters in advance.关键词：RPCL;fuzzy entropy;color image;clustering segmentation3027|182|0更新时间：2024-05-07
- 摘要：A novel algorithm based on dynamic adaptive voxel-growing method is proposed to segment 3D CT pulmonary volumes.It selects the optimal parameters for segmentation by dynamically adjusting statistical information about previous segmented regions.To avoid the disadvantage of leaking during segmentation with the conventional voxel-growing based methods,it adopts a process to mutually utilize segment results between both of lateral lung leaves,which in turn benefit the discriminative segmentation on left and right lung leaves.Compared with conventional algorithms,the accuracy and robustness of segmentation are enhanced,and still this algorithm has the advantage of pure 3D processing ability.Groups of experimental data is verified with this algorithm,including data about the healthy people and types of patients with lung diseases.Results are effective,which imply that this algorithm is potentially valid for future clinical diagnosis applications.关键词：3D medical image segmentation;dynamic voxel-growing;pulmonary CT image3474|195|0更新时间：2024-05-07
- 摘要：Segmentation of brain tissues is very important in medical image analysis.Support Vector Machines(SVM) is considered a good candidate because of its good generalization performance,especially for dataset with small number of samples in high demensional feature space.This paper investigates the segmentation of magnetic resonance brain tissues image based on SVM.Experimental results show that the influence of kernel function and model parameters on the generalization performance of SVM is significant;SVM is suitably used as learning classifier of small sample size;To segment targets with blurry edges,intensity non-uniformity and discontinuity(such as medical images),SVM approach is a good choice.关键词：support vector machines;segmentation;brain tissues2700|187|0更新时间：2024-05-07
- 关键词：MAP-MRF;three-dimensional echocardiographic;3D reconstruction;image restoration2674|210|0更新时间：2024-05-07
- 摘要：The 3D volumetric model of a tooth is constructed based on the image data of CT cross-section layers.Each cross-section image,whose interval is 0.5mm and thickness is 1mm,is treated by two-value conversing,boundary extracting and boundary key points selecting.With the boundary key points of four neighborhood layers,three-dimensional solids between the two adjacent layers are produced according to 3D-Delaunay tetrahedron algorithm.Its validity of each solid element is checked if their internal products between the estimated outward normal vector and the vector from sphere center to vertex are all positive or not.All obtained valid 3D solid elements are combined with Boolean operation,and the volumetric model of a tooth is then achieved at last.This approach is proved to be simple and effective with a tooth sample.关键词：CT slice image;3D volumetric modeling;boundary extraction;3D-Delaunay tetrahedron algorithm2906|176|0更新时间：2024-05-07
- 摘要：Real-time detection algorithm for dim small moving target in image sequences is one of the key algorithms in optical image precise homing guidance systems.At first a brief introduction of current detection algorithms based on temporal profile analysis is addressed,and then noise and target's pixel temporal profile variance is thoroughly studied,finally a detection algorithm based on adaptive recursive variance filter(ARVF) is proposed.The designed algorithm can be used to improve the signal to noise ratio(SNR),so that the point target is easily segmented out.The results of simulated temporal sequences and real image data experiments show that the target(detection's) aim can be fully reached.The proposed algorithm can be implemented recursively and in parallel,which is practical for real-time homing guidance systems.The limitation of the algorithm and future work are also briefly discussed.关键词：dim point target;detection;temporal profile;ARVF2887|217|0更新时间：2024-05-07
- 摘要：Because of high restriction on False Reject Rate,False Accept Rate,computational time cost and template size,fingerprint verification is a challenging task.A novel feature named complex direction feature is proposed in this paper,which consists of ridge rotation direction and four local relative directions.The fingerprint verification consists of two stages: complex direction feature based minutiae alignment,which speeds up alignment;matching score calculation by complex direction feature,which makes matching more robust.Experiments were performed on the fingerprint database used by Fingerprint Verification Competition,and the experimental results show that the proposed algorithm can achieve reliable and fast fingerprint verification.关键词：fingerprint verification;minutiae;complex direction feature2753|154|0更新时间：2024-05-07
- 摘要：Gait is an emergent biometric aimed essentially to recognize people by the way they walk.Gait as a biometric can be seen as advantageous over other forms of biometric identification techniques,for it offers the possibility to identify people at a distance without any interaction or co-operation from the subject.This paper proposes a novel automatic gait recognition method,which extracts gait signature from legs of the subject.For each image sequence,background subtraction based on chromaticity distortion is used to segment moving objects.Boundary tracking algorithm is then used to find perimeter pixels in each processed binary image sequence.This paper makes use of Hough Transform to locally extract the lines which represent legs,and thus obtains inclination angles of upper legs and lower legs.The angles are then fitted to a fifth-order polynomial by least squares method.The polynomial curve is expressed by a Fourier series.The lower-dimensional gait signature vector,that is,the product of phase and magnitude,is derived from phase and magnitude spectra.Fisher Linear Classifier is used to validate the performance of the proposed algorithm on small database samples and the correct classification rate is 79.17%.The recognition rate is still good for these unideal outdoor image sequences.关键词：gait recognition;signature extraction;background subtraction;Hough transform3041|162|0更新时间：2024-05-07
- 摘要：Defects on the surface of steel strips are main factors to evaluate quality of steel strips,and surface inspection is of great importance to improve quality of steel strips.Traditional surface inspection by human inspectors is far from satisfactory.In this paper,an approach to detect real-time surface defects of steel strips based on feed-forward neural network(FFN) is discussed.The experiments show that the method is effective.关键词：defect detection;neural network;cold rolled strips;image processing;machine vision2610|227|0更新时间：2024-05-07
- 摘要：Traditional subpixel edge location algorithms use local model to realize precise edge location.Which leads to side effects such as Zigzag effect and discontinuous edges.We present a new subpixel edge location algorithm based on partial differential equation(PDE),where level-set reconstruction(LSR) is introduced to smooth edge along their tangents.Anchor and topology constraints are used to avoid over-smoothing and keep edge topology structures.Subpixel edge location experiments with both geometric and natural image show that this method can remove zigzag effect while keeping edge smoothness and continuity.关键词：subpixel;edge location;level set reconstruction2859|188|0更新时间：2024-05-07
- 摘要：The commonly used reconstruction method from images acquired by a calibrated camera but with unknownmotion parameters is the essential matrix based one in the literature,i.e.,from the essential matrix to decompose therotation matrix and translation vector at first,then to reconstruct the scene by the standard stereo method.A less well-known and conceptually rather more abstract method is the so-called absolute dual quadric based method.However,basedon extensive experiments on simulated data as well as on real images,we showed that the absolute dual quadric basedmethod consistently outperforms the essential matrix based one in terms of reconstruction accuracy and robustness,hence itis highly recommended to use the absolute dual quadric based method rather than the commonly used essential matrix basedone in practice.关键词：projective reconstruction;metric reconstruction;essential matrix;the absolute dual quadric2553|173|0更新时间：2024-05-07
- Postal code：100190
- Tel：010-58887035/58887030/58887418 Email：jig@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备05080539号-4
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0