最新刊期
卷 18 ， 期 7 ， 2013
- 摘要：The image denoising and edge enhancement based on parabolic PDEs are the hot topics, but the existing diffusion equations have always made the edge smooth and hence the images become blurred. Based on the anisotropy diffusion, an equalized net diffusion model, with a modulator and a constraint term, is presented in order to not only denoise the image but also enhance the object edge and the image structure. An explicit calculation scheme is desired and the model is tested subsequently. The experimental results show that our present method is superior to the existing methods in the processing effect above.关键词：image denoising;edge enhancement;parabolic equation;equalized net diffusion3194|377|0更新时间：2024-05-07
- 摘要：For images without obvious lines, a novel skew correction method based on feature points matching is proposed. In this method, the correction does not use actual lines, but relies on the lines constructed by the feature points extracted from the images. It is translation and scale invariant. Through bidirectional best correlative coefficients calculation, a high matching rate can be achieved. Mismatching can be decreased, by using the method based on the Law of Large Numbers. It can get the correct slant angles by using only two matching pairs. In addition, due to the special application background of this paper, a disk template is designed to extract the Eigenvector of the features, which leads to rotation invariance.关键词：feature points matching;skew correction;bidirectional best correlative coefficients;law of large numbers3081|307|0更新时间：2024-05-07
- 摘要：In this paper,we are aiming at random-valued impulse noise detection in two boundaries, using advanced boundary discriminative noise detection(ABDND)and a global histogram to obtain the noise boundary. Although we get good detection results, the rate of false detection increases for ABDND when the range of the noise boundary increases. A modification of the ABDND(MABDND)is therefore proposed in this paper. It includes two stages. First, it uses the global histogram to obtain the noise boundary identical to the ABDND. Second, it uses the statistic of a part of the histogram to find false detected pixels in the first stage, and marks them as uncorrupt pixels. The merit of MABDND is to use the confirmation technique in the second stage to rectify many false detectied pixels in the first stage to keep a low rate for both miss detection and false detection. Image Lena and peppers are used for simulations, and the experimental results show the performance of MABDND is better than that of ABDND, especially, when the range of random-valued is wide MABDND is more robust.关键词：random-valued impulse noise;boundary discriminative noise detection(BDND);advanced boundary discriminative noise detection(ABDND);switching median filter;noise detection2922|308|0更新时间：2024-05-07
- 摘要：The traditional total variation (TV) models based on local operators for color image denoising has in some problems, such as smeared edges, smeared textures, staircase effects, and mosaic effects. In this paper, the Tikhonov model, TV model, multi-channel total variation (MTV) model, color total variation (CTV)model based on local operators are extended to the non-local color Tikhonov (NL-CT)model, non-local layered total variation (NL-LTV)model, non-local multi-channel total variation (NL-MTV)model, non-local color total variation(NL-CTV)model via non-local operators for color texture image denoising. Using auxiliary variables and the Bregman iterartive parameters, we design their fast Split-Bregman algorithms. Experiments show that all of them solve the above mentioned effects and demonstrate excellent properties of edge preserving, texture preserving, and smoothness preserving. NL-CTV has the best result,but its computational efficiency is low.关键词：color texture images;image restoration;non-local TV models;Split Bregman algorithm;feature preserving3664|387|0更新时间：2024-05-07
- 摘要：The similarity of one image plays an important role in super-resolution reconstruction. However, a single image can not provide sufficient similarity information. Therefore, a video super-resolution reconstruction algorithm based on similarity constraints is proposed. Making initial motion estimation with flow field and fine block matching, the similar information, which can be used to modify the reconstruction error in iterative process of iterative back projection, will be extracted from video sequences. The experimental results show that the proposed algorithm has improve a the image details, moreover, it can remove the artifacts and obtain clear and sharp edges in visual perception.关键词：similarity;back projection;super-resolution;video2658|501|0更新时间：2024-05-07
- 摘要：Estimation of the fingerprint orientation field plays an important role in automatic fingerprint recognition systems. In this paper, a new algorithm for computing the fingerprint orientation field is proposed. The basic idea behind the algorithm is to estimate the local region orientation. We use linear projection analysis and weighted lower order two-dimensional Legendre polynomials in the sine domain. This method helps to accurately extract low quality fingerprint image orientation fields. Experiments on the database of FVC 2000 show that, compared with the previously proposed gradient-based methods, the proposed method is more accurate and more robust against noise, and is able to better estimate fingerprint orientation field of low quality fingerprint image with large noisy areas.关键词：fingerprint recognition;linear projection analysis;Legendre polynomials;orientation field3518|411|0更新时间：2024-05-07
- 摘要：The basic idea of identifying the direction of motion blurring by directional derivatives is that the original image is an isotropic first-order Markov random process. However, the real result are not good. The main reason is that the images do not meet this assumption strictly. Judging from the entire image, the shape of the objects, flat areas, and some other factors tend to weaken the physical premise, and cause inaccurate calculation result of the differential image. In this paper, we propose an optimization method. First, multiple feature blocks are extracted via the image local variance, and then the statistical results of the motion blur direction identification from every block are retrieved. The experimental results show that our method can still obtain good identification accuracy even when the error of the identification of traditional directional derivative method is big.关键词：directional derivative;Markov random process;local variance;weighted average method;blocked statistics method2559|218|0更新时间：2024-05-07
- 摘要：In order to solve the problem of deformable objects segmentation with a fixed shape prior, a shape prior constrained and parameter adaption level set segmentation method based on kernel principal component analysis(KPCA)is proposed. First, the KPCA method is used to get the base vectors in the shape prior feature space. Then, the Parzen window method is used to estimate the results of the original image for image data term and an affine transformation is performed to align the image region of interest and prior shape training set to add shape priors to the segmentation model. At last, a parameter adaptive method is introduced when solving the evolution equation based on level set method. Experimental results show that our method can effectively segment objects with different attitudes in comparison with the Chan-Vese (CV) model and single prior shape constrained level set methods.关键词：image segmentation;shape prior;kernel principal component analysis(KCPA);affine transformation;parameter adaption2560|796|0更新时间：2024-05-07
- 摘要：Subtractive clustering based methods have been well known for data clustering problems. However, due to the computational demands of these approaches, clustering for large scale datasets, such as spatio-temporal data and images, have been slow to appear. A novel subtractive clustering method based on Nyström approximation is proposed. The proposed method is based on the famous Nyström method. Combined with the density value computation characteristics for each sample of the classical subtractive clustering method, we apply Nyström theory to approximate the density value for each data point which has not been sampled. Finally, we complete the whole clustering procedure using classical subtractive clustering method in modifying the density values in each circulation. The proposed method substantially reduces the computational requirements of subtractive clustering based algorithms, making it feasible to use subtractive clustering to large scale subtractive clustering problems. Density value of samples could be approximated quickly using only a small number of samples. The experiment results on artificial datasets, color images, and UCI machine learning repository show efficiency in comparing with classical subtractive clustering method.关键词：subtractive clustering;Nyström;density value approximation;time complexity2960|285|0更新时间：2024-05-07
- 摘要：Regarding the numerous stellar targets with similar image characteristics that interfere with space targets detection in fixed-direction terrestrial astronomical opto-electronic observation systems, an algorithm of stellar targets suppression based on temporal-spatial domain fusion is proposed. In the time domain, the conventional image difference algorithm is improved and the dilative difference method is used to suppress stellar targets. In the space domain, mathematical morphological reconstruction is used to restore the full image of the detected stellar targets, and then stellar targets are suppressed through the difference between images before and after the morphological reconstruction. Based on different temporal-spatial domain fusion mechanics, two different temporal-spatial fusion methods are proposed to suppress stellar targets. Through comparison of real-measured data, the results show that the algorithm presented in this paper achieves a good stellar targets suppressing effect: 1)stellar targets and their edge pixels are efficiently suppressed; 2)the faint stellar targets are suppressed well; 3)signal-to-noise ratio of space targets are improved after the suppression of stellar targets.关键词：stellar targets suppression;space targets detection;temporal-spatial fusion;morphological reconstruction;dilative difference2871|213|0更新时间：2024-05-07
- 摘要：A novel feature representation method based on super-interest points is proposed and applied to human action recognition. Interest point features describe the salient change of local point information when humans are in the state of movement, however, the drawback of these features is that the rich spatio-temporal relationships among the discrete interest points are not being used. According to the spatio-temporal distance of interest points, a broad first search neighbors algorithm is used to cluster these adjacent points into super-interest point. The super-interest point features reflect the variety of the human limb in a certainly spatio-temporal scope. Compared to the existing interest point method, the spatial and temporal information among the points, are all included in our super-interest points, making the features more discriminative. Finally, a two hierarchical classifier recognizes the human actions using super-interest point features. The experimental results show that our method achieves a good recognition rate.关键词：super-interest points;action recognition;broad first search neighbors;spatio-temporal distance3119|307|0更新时间：2024-05-07
- 摘要：Using contour grouping to get perceptual objects from natural images is a traditional difficult problem in computer vision. To provide a cue combination model, which accords with human visual perception and the statistical of the natural images, is the key problem of the definition of grouping cues and the improvement of grouping quality. According to the conditions when a linear cue combination rule works, a cue combination model for contour grouping, which fits the ecological joint distributions of continuity and similarity in Gestalt cues, is proposed. The cue combination model helps to explain how two independent cue channels could give rise to the particular form of the joint distribution of two correlated cues. It also overcomes the cue combination problem between correlated cues which can't be resolved by discriminative model. The proposed model is a quantitative model of Gestalt cues which accords with the statistical of the natural images and human visual perception better.关键词：contour grouping;Gestalt rules;cue combination;neural computing2948|239|0更新时间：2024-05-07
- 摘要：To deal with different road conditions and environment situations, a hierarchical vanishing point detection method based on holistic image features is proposed. Four road types are classified based on an abstracted holistic road feature, which also coarsely locates the road. After classification, the lane or edges in a specific scale is chosen for the line-segment-detection or region segmentation, and for the voting of the intersection point. Additionally, an inverse projective image or region entropy verifies the result. Interference of vehicles and shadows are eliminated during image pre-processing to further improved the accuracy. Our results show that the hierarchical method can choose the right road type and robustly detect the vanishing point under complex road scenes in real time, achieving an improvement for accuracy and efficiency at 37.5% and 20% compared to the state-of-the-art method based on intrinsic line orientation and color texture properties.关键词：complex road scenes;road type estimation;holistic image feature;vanishing point extraction and verification2700|609|0更新时间：2024-05-07
- 摘要：The binding problem is regarded as one of the central themes of cognitive science and neuroscience. In order to solve the binding problem of shapes and colors for color images, an advanced vector model for the binding problem, based on the simplified pulse coupled neural network(PCNN), is presented in this paper. In this model, the first pulse emission time of a neuron separates the different characteristics, while the input stimulus of the neurons itself binds the different characteristics. The experimental results indicate that the model can achieve good feature separation as well as feature binding and it discovers the optimal iteration times. The method has certain reference value for the research and application of the binding problem of color image.关键词：color image;feature binding;pulse coupled neural network;vector model;automatic determination2741|232|0更新时间：2024-05-07
- 摘要：Traditionally, the non-photorealistic rendering pyrography technology has the problems of a low rendering speed and of a single background; therefore details cannot be reflected in the final artistic effect. In this paper, a new measurement model about pyrography simulation based on the deviation mapping is proposed to avoid these disadvantages. First, the background of the pyrography could be obtained based on the texture synthesis technique. Many optimizations such as searching along a spiral path, edge expansion, L2 distance computing predigesting accelerated synthesis speed and the final results is also advanced. Second, based on the tensor field, the foreground image is filtered based on the anisotropic and line integral convolution methods to show enjoyable style features. The technique of image enhancement is also used to enhance the local structure and details. Additionally, the Phong illumination model is used to map the foreground image to the background image. Experimental results show that the details and freehand enjoyable character can be produced. So, the proposed method can simulate the real pyrography effect.关键词：non-photorealistic rendering;pyrography;texture synthesis;deviation mapping2859|328|0更新时间：2024-05-07
- 摘要：When abnormal conditions occur in point clouds, the normal adjustment may have erroneous results when using the minimum spanning tree algorithm, while the efficiency is low when using the surface reconstruction algorithm. In order to solve this problem, an optimal algorithm for normal adjustment of point cloud is proposed. It deals with three abnormal conditions separately. For the thin feature condition, it exacts feature points and reverses orientations compulsively. For the perpendicular normal condition, the neighboring region is expanded to get the tendency of the normal. For close-by surfaces condition, ambiguous neighbors are removed from the K-nearest neighbors to avoid creating an erroneous minimum spanning tree edge. Experiments show that the algorithm can adjust the normals correctty even when such abnormal conditions exist. Compared with surface reconstruction algorithm, the algorithm can adjust the normals more efficiently.关键词：point cloud;normal adjustment;minimum spanning tree;K-nearest neighbors;surface reconstruction2885|278|0更新时间：2024-05-07
- 摘要：Fourier imaging in MRI application has the dilemma that using extended k-space sampling to improve image resolution also degrades the signal-to-noise ratio(SNR)because of the Fourier uncertainty. In this paper, we propose a new method using anisotropically constrained image reconstruction based on a penalized maximum likelihood optimality model, which is an optimization problem instead of a discrete Fourier transform (DFT) approach. Anisotropic regularization for enforcing anatomical prior information is proposed, where directional regularization operators apply to the smooth areas, neighboring edge areas and edges respectively. Experimental results show that the proposed method enables extended k-space sampling while suppressing Gaussian noise and reducing the reblurring problem and the Gibbs ringing artifacts of existing constrained reconstruction methods.关键词：anisotropically regularization;constrained image reconstruction;penalized maximum likelihood optimality;anatomical prior2445|238|0更新时间：2024-05-07
- 摘要：To automatically detect hard exudates from fundus images, and to develop an automatic diabetic retinopathy screening system, an automatically detecting approach based on RBF neural network and threshold segmentation was established and studied. First, the green channel of the original fundus image is coarsely segmented by an improved Otsu thresholding based on minimum inner-cluster variance, and candidate regions are obtained. Second, several features of candidate regions are extracted and selected by means of logistic regression. Finally, the RBF neural network is built with the optimal subset of features and judgments of these candidate regions. Furthermore, post-processing is carried out to improve the detection accuracy. The approach is tested on a new set which contained 50 fundus images with variable color and brightness. With an image-based criterion, sensitivity of 100%, specificity of 90.9%, and accuracy of 96.0% are achieved. Average sensitivity of 93.9% and average positive predict value of 95.5% are also achieved with a lesion-based criterion. Furthermore, the average time cost in processing an image is 13.6 s. Results suggest that the approach is stable and reliable, and can fast and effectively detect hard exudates from fundus images.关键词：fundus images;hard exudates;RBF neural network;threshold segmentation;automatic detection2374|256|0更新时间：2024-05-07
- 摘要：Currently, related image quality assessment methods can not appropriately measure structure-preserving capability of synthetic aperture radar (SAR) image despeckling, to solve this problem a no-reference improved equivalent number of looks of ratio image(IENLR) assessment method is put forward. To begin with, related image quality assessment methods are reviewed, where their weaknesses to measure structure-preserving capability are investigated. Furthermore, from equivalent number of looks of ratio(ENLR) image as a point of departure, the computing formula of IENLR is constructed, meanwhile theory analyses prove that IENLR is a basically monotone, bounded function with regard to structure-preserving capability, and that IENLR is adjustable in the light of its parameter. Finally, simulated experiments further verify that IENLR is excellent and superior to other related image quality assessment methods from the viewpoint of measuring structure-preserving capability.关键词：synthetic aperture radar(SAR)image;speckle noise;image despeckling;structure-preserving capability2619|264|0更新时间：2024-05-07
- 摘要：Building contours and corners, as common registration primitives of multi-platform LiDAR data, are drawing more and more attraction on their extraction. Density of projected points(DoPP)method is a usual approach for building contours and corners extraction from terrestrial LiDAR, however previous studies paid little attention on the determination of grid density threshold, which directly affects the extraction results.In this paper,we propose a building contour density estimation method, which can provide a grid density threshold according to the actual point cloud, so that accurate building contours can be extracted. On this basis, the extracted contours are segmented and recovered using contour segmentation and density extension method, thus complete building contours are obtained. Finally, building corners are gained through the intersection of building contours. The experiment shows that the proposed method can effectively extract building contours and corners from terrestrial LiDAR with high correctness, completeness and positioning precision.关键词：terrestrial LiDAR;building contours and corners;grid density threshold;contour density estimation method3677|285|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