最新刊期
卷 14 ， 期 8 ， 2009
- 摘要：Linear array push broom sensors have been widely researched and applied in the aerospace area and remote sensing applications, and have started a new approach for photogrammetry and remote sensing application Research on extrapolation of attitude parameters of push broom satellite with less ground points have become a new research It is very important to target objects in a region where it is difficult to obtain ground control points According to the characteristics of multiple central perspective imaging,we built the sensors model based on collinear equationsWe solve orientation parameters of different scan lines by space resection When the extrapolative depth is large, we take a single ground control point as the constrains The experiment data is six Spot images of the same orbit in our country We analyzed the extrapolative results: when the extrapolative depth is 240 km, the mean error is about 1 km, when extrapolation considers a single ground control point, the result is better than the former and the mean error is below 1 km We can improve our results in the future by collecting other related information for the extrapolative image关键词：linear array push broom sensors;collinear equation;extrapolation;few ground control point5675|290|0更新时间：2024-05-07
- 摘要：This paper mainly suggested two ways of matching method in spatial domain for high-resolution remote sensing images in order to obtain equalized distributing GCPs Ordinary matching method first manually acquires a few initial GCPs, then extracts GCPs with feature detection operators The final step is the template matching This paper suggested equalizing the GCPs either during the feature detecting or after the template matching At the end of the paper, experiments proved the effectiveness in accuracy improvement with the proposed methods关键词：remote sensing images;image matching;GCP equalization6355|308|0更新时间：2024-05-07
- 摘要：To investigate the fusion application based on multi-band wavelet transformation for remote sensing imagery, this paper presents an introduction to m-band wavelet transformation theory and a typical case in image fusion Meanwhile, results from dyadic wavelet, dyadic wavelet packet and m-band wavelet transformation are compared; and the experimental results from different image fusion schemes are compared, including substitutive, addictive, as well as weighted models It has been found that, m-band wavelet-based schemes perform better than dyadic schemes In a word, it is the only scheme, not the best scheme That is to say, the results from wavelet-based fusion methods can be improved by choosing more appropriate wavelet bases and more sophisticated models for injecting detailed information; however, there is still a long way to go关键词：BJ-1 satellite image;image fusion in pixel level;multi-band wavelet transformation;comparison of fusion schemes5151|323|0更新时间：2024-05-07
- 摘要：Integrating structural similarity and global quality measurement index, a new non-reference fusion quality assessment method for multispectral remote sensing imageries is proposed. First, the spectral quality is obtained from local spectral quality based on the weighting strategy of mean variable. Then, the spatial quality is achieved from local spatial quality based on the weighting strategy of standard deviation. Finally, a combined quality is calculated by linear addition. Validating experiment is carried out on eight common fusion methods using CBERS-02 multispectral and Cartosat-1 panchromatic images. The result shows the new method is consistent with the judgments based on human visual system, which indicates its efficiency.关键词：fusion;quality assessment;structural similarity;global quality measurement index5399|358|0更新时间：2024-05-07
- 摘要：Moment techniques are an important research area and have been widely used in image processing and analysis such as pattern recognition, pattern matching, image compression, digital watermarking and so on For helping the researche rs to have a general understanding in this technique this paper presents a survey of recent and traditional moment techniques in image analysis It discusses various types of moments, especially some fundamental aspects of discrete orthogonal moments, such as computation capacity, reconstruction accuracy, and invariance etc in detail Open issues and possible future research direction are also pointed out关键词：moment function;moment invariants;image analysis3945|299|0更新时间：2024-05-07
- 摘要：In order to make people have a broad understanding to the application of medical image deformable registration based on models in image guided radiotherapy, The recent publications that research the application of deformable registration in image guided radiotherapy are concisely classified and summed up according to the three major steps of deformable registration: the selection of deformable model and feature space, the selection of similarity metrics and space optimization, as well as the verification of deformable registrationThe statistics of classification shows that the research directions of deformable image registration which basing to ensure the accuracy and precision of registration, are to combine different kinds of methods, such as multi-resolution, multilevel deformable registration, using parallel computing, and the combination of different kinds of optimization algorithms and so on, in order to shorten the registration time and to meet the clinical demands关键词：deformable registration;deformable model4094|273|0更新时间：2024-05-07
- 摘要：Due to the lack of population diversity,the simple genetic algorithm often leads to the premature convergence and affects restoration results when applying the algorithm in super-resolution image restoration.A catastrophis genetic algorithm for super-resolution image restoration is proposed which takes the fitness value standard deviation as the catastrophic condition and the conjuate gradient algorithm as the catastrophic operatro.The operator recreates the population and improves the population diversity.After analyzing the catastrophic condition and the catastrophic scale,the method to adjust catastrophic condition according to the number of genrations is proposed.The experiments demonstrated the effectiveness of the algorithm.关键词：super-resolution;image restoration;catastrophic genetic algorithm(CGA);population diversity3835|282|0更新时间：2024-05-07
- 摘要：This paper discusses a watermarking method towards keeping robustness and high accuracy of vector graphics. The graphics is treats as a set of curves in the embedding algorithm, and each curve is divided into several shorter curves under a given threshold. And then a watermark bit combined with user certificate is embedded into each segment around the node for every cutting curve. To extract the watermark, it also divides up the watermarked graphics according to the division of the original graphics. Then, it extracts the watermark bit from each curve with the inverse procedure of the embedding process. Finally, the method calculates the similarity between the original watermark bits and the extracted ones from the watermarked graphics, and determines whether the watermark exists or not. The experiment results showed that the method not only guarantees the accuracy of vector graphics but also prossesses good robustness, such as it gives 100 similarities under no attack or only geometric transformation with the map; the anti-cropping ability gives 0.9 even for the graphics cropped 2/3. And it also has the ability of anti-compression of D-P approach and the experiment curve of the similarity threshold given which helps to control the anti- attack ability of watermark and set parameters for an automatic procedure of watermark detection.关键词：vector graphics;digital watermark;robustness;similarity;curve partition3944|240|0更新时间：2024-05-07
- 摘要：Mathematical filters based on mathematical morphology is a non-linear filter. Based on the erosion and dilation operations a class of generalized open and close morphological filters is constructed by using different structuring elements in this paper. These filters possess some important properties such as translation invariance,duality,increasing and so on. In order to verify its effectiveness，we simulate against different noise with generalized open and close morphological filters. The results of computer simulation show that these filters can preserve the geometrical features and edges of the image when they suppress noise in the image and their speed is faster than the traditional morphological filters. They possess calculation simplicity and easy to be real-time processed in parallel.关键词：image processing;morphological filters;morphological trsnsform;erosion;dilation3577|227|0更新时间：2024-05-07
- 摘要：In this paper， we propose an enhanced cross-diamond-hexagonal search algorithm to solve the block-matching motion-estimation problem in video coding. This algorithm is an improvement over CDHS， eliminating some checking points of CDHS algorithms. Experimental results show that the proposed method performs faster than the CDHS whereas similar quality is preserved.关键词：video encoding;block-matching motion estimation;cross-diamond-hexagonal search3990|225|0更新时间：2024-05-07
- 摘要：Compared with traditional triangle meshes, point models are suitable to represent irregular objects with complex geometry However, there are few digital watermarks algorithm based on point models In order to protect effectively 3D point model, a new method for digital watermarking of a 3D model defined as a set of un-oriented points based on geometrical property is proposed in this paper Our approach is to modify the distance between each point and the center of the model after principle component analysis according to the watermarking bits At last, embed watermarking by shifting the mean value of the distance distribution. Mesh pre-processing doesn’t need before watermarking is embedded into the point model in the algorithm. In addition, geometrical distortion degree between the original and the tested model is analyzed, and two parameters of watermark strength factor and the size of bin that affect the robustness of the proposed methods are discussed. After conducting the affine transformation and vertex reordering attacks, the original model is not needed during the watermarking extraction Experiments show that this approach is greatly robust against attacks, such as affine transformation, resorting, simplification, cropping, nosing and various distortion ones with local geometric alterations.关键词：point model;geometry property;watermarking;geometry distortion;robustness3520|286|0更新时间：2024-05-07
- 摘要：So far information hiding algorithm based on DCT domain is the most extensive one in all transformation domain algorithm applications Its most serious disadvantage is the low capacity for information hiding The reason why only small capacity information can be hidden in DCT domain of an image is analyzed, and a high capacity blind information hiding algorithm based on DCT domain is proposed In our algorithm, the uniform spectrum processing is first applied to the original image Then, a global DCT transformation of the image is finished Finally, the information, which is to be hidden, is embedded into the low frequency coefficients in DCT domain Moreover, the information hiding is implemented by using segment quantification method, extracting the hidden information without using the original image Experiment shows that our algorithm has a high capacity of information hiding and good imperceptibility关键词：information hiding;DCT;uniform spectrum;quantification;blind extracting4147|342|0更新时间：2024-05-07
- 摘要：Through analysis shortcoming of the ROF(Rudin, Osher and Fatemi) model and LLT（Lysaker,Lundervold and Tai) model in denoising processing, and the difference between texture and noise, combining decomposition model, TV-norm and fourth-order PDE, the article proposes the DD(decomposition and denoising) modelWhen processing noise image, the new model decompose an noisy image into three parts, structure, texture and noise, and thas achieves denoising and decomposition Further through the experiments,we testify rationality and validity of the DD model and the algorithm关键词：image decomposition;image denoising;structure;texture;noise3801|262|0更新时间：2024-05-07
- 摘要：AVS is a fully intellectual property of second generation source coding and decoding standard owned by China And AVS is a set of standard system which includes system, video, audio, and media copyright management It will be potentially and widely used in digital video and audio industry of China This paper designs and implements an AVS-based video surveillance system, which can automatically process AVS video streams In the AVS-based video surveillance system, the AVS coding and decoding standards are combined with other processing techniques, such as background/non-background classification, face detection and recognition, and two-level indexing structure to automatically identify the persons who appeared in the surveillance videos The videos can be retrieved quickly by inputting texts, pictures, and video segments The AVS-based video surveillance system can be used in many areas such as digital surveillance, access control, and etc关键词：AVS-based;video surveillance system;automation;background/non-background classification;face detection and recognition;indexing structure3316|227|0更新时间：2024-05-07
- 摘要：Objective image quality assessment, which automatically values the image quality consistently with subjective perception, is important to numerous image and video processing applications, especially in image coding In this paper, a new objective measure for image coding quality based on image structure is proposed First of all, the structure information is calculated by projecting the reference image onto the base functions using the algorithm of matching pursuit Secondly, the structure information of the coded image in the order of importance is computed and gained, and the characteristics of the importance of the structure information are given, using the structure information of the reference image Finally the objective quality score is concluded by comparing the differences of structure information between the reference and the coded images And the separable 2-D Gabor base functions are selected as the dictionary for the algorithm of matching pursuit Experimental results show that the proposed approach is consistent with the subjective quality score关键词：image quality assessment;matching pursuit;image structure;image coding4274|232|0更新时间：2024-05-07
- 摘要：The relation of Fresnel diffraction and factional Fourier transform are analyzed, a fast simulated algorithm of fractional Fourier transform based on FFT is proposed By making use of fractional Fourier transform as the center tool, the new computing methods of holograms are obtained and the computer simulation holographic generation and digital reconstruction are realized By introducing the spatial light modulators DMD, a holography imaging system to complement the holography display is set up关键词：holography display;fractional Fourier transform;digital micromirror device (DMD)3690|275|0更新时间：2024-05-07
- 摘要：The design of a real-time video compression and coding system based on wavelet is put forward for realizing the effective real-time video compression The key technical point in this system is wavelet transform Firstly,the device ADV612 is used for real-time hardware video compression and coding, then DSP TMS320VC5509, video data which is 760×420 resolution and 25 frames per system meet 512 kbps~15 Mbps and has a number of advantages of high real-time speed, simple system structure and good extensibility关键词：wavelet transform;video compression and coding;digital signal processing（DSP）;ADV6123281|236|0更新时间：2024-05-07
- 摘要：Weak and small target detection is a crucial problem in infrared technology Facing the problems of losing weak/small target in single threshold segmentation and insufficient energy accumulation for fast-moving targets in the target detection of infrared image sequence, a new method based on multi-threshold classification and backward verification has been proposed in this paper After background suppression, an adaptive multi-threshold classification is adopted for the extraction of multi-class candidates, which has enhanced the capability of weak and small targets extraction When the candidate in the current frame can not be justified due to its weakness and movement, a kind of spatial-temporal pipeline will be constructed according to the candidate’s movement in adjacent frames, and a backward search will be carried out for possible existence of any kinds of candidate After that, the energies of these candidates will be weighted accumulated with that of the candidate in the current frame and justified with a fixed energy threshold In the end of this paper, some experimental and comparison results using real infrared image sequences will be given to show the effectiveness of this new method关键词：infrared target detection;weak and small target;fast-moving target;multi-threshold classification;backward verification3611|225|0更新时间：2024-05-07
- 摘要：An improved algorithm based on randomized circle detection is presented The large amount of useless accumulations yielded by random sampling of randomized circle detection algorithm and the useless square computations resulted from the outlier edge points during the evidence-collecting phase are solved First, the edge pixels with 8-adjacency connectivity are connected, and three edge pixels in the same connected curve which can exactly determine a possible circle with center and radius are randomly picked When a possible circle is found, an evidence-collecting process is implemented to further verify whether the possible circle is a true circle or not, in which only the edge points locate between the circumscribed square and the inscribed square of the possible circle are considered Compare to conventional detecting algorithm, the proposed algorithm consumes less computing resources, has better detection performance when comparing to other methods and can detect circular arc or partly overlapped circles by the experiments关键词：randomized circle detection;8-adjacency connectivity;multi-circles detection3534|237|0更新时间：2024-05-07
- 摘要：As a critical step in many multimedia applications, shot boundary detection attracts many research interests in recent yearsMost present methods measure the similarity among video frames based on its low-level feathers However, they are sensitive to the change in brightness, color, motion of object,camera motions and the quality of video This paper proposes an innovative shot boundary detection method for news video based on image segmentation and object tracking It combines three main techniques, namely, the partitioned histogram comparison method, the image segmentation based on wavelet analysis and the object tracking The partitioned histogram comparison is used as the first filter to effectively reduce the number of video frames which need segmentation and object tracking The unsupervised image segmentation based on wavelet analysis and object tracking is robust to those problems mentioned above The efficacy of the proposed method is extensively tested with more than 3 hours of CCTV and CNN news programs, and that 964% recall with 972% precision has been achieved关键词：shot boundary detection;image segmentation;object tracking;partitioned histogram3335|246|0更新时间：2024-05-07
- 摘要：Gradient correlation matrixes representing local geometrical features of planar contour is constructed, and properties of eigenvalues and eigenvectors of GCM are analyzed and proved based on Lagrange multiplier optimal method and Γ corner model Thus, the corresponding geometrical meaning is obtained The determinant of GCM is defined as corner response function and the corresponding corner detection algorithm is presented Finally, a number of experiments demonstrate that the proposed algorithm has a good detection performance and robustness to the various geometrical transformations and noise关键词：corner detection;gradient correlation matrix;planar curve;determinant3837|262|0更新时间：2024-05-07
- 摘要：Principal curves is a new feature extraction method based on nonlinear transformation They are smooth self-consistent curves that passes through the “middle” of the distribution They perfectly reflect the structural features of the data The paper chooses principal curves to extract strokes of characters and segments numeral strings by grouping strokes based on the confidence of the classifiers The classifiers based on the segmented contour feature and the normalized template features are combined and experimental results indicate that the correlation of these two features is small The paper modifies the confidence of the combined classifier by posterior probabilities which are estimated by a novel class-conditional confidence transformation approach Experimental results indicate that the method is effective in the segmentation of numeral strings关键词：principal curve;multi-classifier combination;confidence transformation;posterior possibility;segmentation of numeral strings;stroke grouping3494|257|0更新时间：2024-05-07
- 摘要：In this paper, a facial expression recognition method based on feature fusion and fuzzy kernel discriminant analysis (FKDA)is proposed This method firstly locates 34 landmark points from each facial image as the Geometric features of the facial image Then, these landmark points are converted into a labeled graph (LG)vector using the Gabor wavelet transformation method， and the LG vector are used as the Gabor feature vector of the facial image Both Geometric feature and Gabor feature are further fused using the canonical correlation analysis (CCA)as the final input facial features for recognition The FKDA method is finally used to further extract the discriminative expression features for classification and the nearest neighbor classifier is used to this goal Experiments on both Japanese Female Facial Expression (JAFFE)database and the Ekman’s ‘Pictures of Facial Affect’ database demonstrate the better performance of the proposed method关键词：fuzzy kernel discriminant analysis;kernel discriminant analysis;facial expression recognition3865|248|0更新时间：2024-05-07
- 摘要：Recently, vision sensor based vehicle detection becomes an attractive research area for the driver assistance system (DAS). But most research was carried out in the day time with a good lighting condition and what little research so far done in the night time assumed no interference of light. Hence a headlights based vehicle detection algorithm at nighttime environment is proposed in this paper. The proposed algorithm detects the rear vehicles via real-time image sequence capture Firstly, bright objects are extracted from the nighttime road scene images Then the extracted bright objects are verified based on rules to eliminate the interferential light sources and obtain the actual vehicle headlights Finally, a knowledge-based method is used to cluster vehicle headlights The clustered headlights are then validated Experimental results demonstrate the feasibility, effectiveness and robustness of the proposed algorithm on vehicle detection at night关键词：vehicle detection;night scenes;vehicle headlights;image processing;driver assistance system3818|254|0更新时间：2024-05-07
- 摘要：A new approach to gait recognition based on fusion of the information of global silhouette and local joint angle is proposed The vector data scanned from horizon, vertical and diagonal of the outer contour of binarized silhouette of a walking person are chosen as the basic image feature Two independent global classifiers are established respectively by the decomposed feature based on the discrete wavelet transformation（DWT) and the nonlinear components of basic gait features extracted based on kernel principal component analysis（KPCA).The coax and knee joint of moving body are simply modeled.The acquired joint angle information is expanded in Fourier series form in view of the periodic character of gait activity. The genetic algorithm is applied to search for the expanding coefficients, and the local feature classifier is established by the normalized eigenvector about joint angle At last, the global and local features are fused based on different Bayesian combination rules on decision level to improve the performance of both identification and verification. This algorithm is applied to CMU database.Extensive experimental results demonstrate that the proposed algorithm performs nicer classification and verification capability关键词：gait recognition;multiple classifiers fusion;DWT;KPCA;arthrosis-based modeling3858|263|0更新时间：2024-05-07
- 摘要：A new algorithm is proposed to segment color image automatically using the intersecting cortical model with the reference of traditional color image segmentation method. The algorithm converts images into HSV color space and selects one of the H, S, and V components with the decision rule of maximum entropy. And it has increased the processing speed greatly compared with traditional color image segmentation method which deals with the three components individually and then merges the results. Our new algorithm costs 257 s, while about one-third of that of the traditional color image segmentation method uses 7533 s. The automatic segmentation with less artificial sets and high accuracy is realized by introducing the maximum cross entropy decision rule into the intersecting cortical model. The new algorithm was compared to the image segmentation method based on max entropy. And the simulation results show that the new algorithm has good performance in color image automatic segmentation.关键词：color image;color space;intersecting cortical model;entropy;cross entropy;image segmentation3599|238|0更新时间：2024-05-07
- 摘要：A three-dimensional representation of vasculature system can be extremely important in image-guided neurosurgery, pre-surgical planning. In this paper, a multi-attribute based spatial continuity fuzzy clustering algorithm (multi-attribute based spatial continuity fuzzy clustering algorithm, MASCFCM) is proposed for segmenting entire blood vessels from the time of flight magnetic resonance angiography (TOF MRA) images. This clustering method takes both the intensity information and the geometrical information into account, while most of the current clustering methods only deal with the former. In this method, a new dissimilarity method, which integrates the intensity and the geometry shape dissimilarity, is introduced. Because of the presence of the geometrical information, the new measure is able to differentiate the pixels with similar intensity values within different geometrical shape structures. To evaluate the algorism, the algorithm is exerted on both 2D and 3D images and the experimental results show that the new algorithm can achieve better segmentation results.关键词：fuzzy clustering;scale space analysis;spatial continuity;blood vessel segmentation.3636|251|0更新时间：2024-05-07
- 摘要：In this paper, we propose a novel video key-frame extraction method based on visual attention model Firstly, the spatiotemporal saliency levels are generated and fused in a motion priority fashion to produce the overall attention degree Then, a new adaptive key-frame extraction algorithm using attention and the variation of attention is put forward For the shot level, the frames with higher attention value are selected as the candidates of the key-frames For the clip level, the key-frame number is generated by the attention variation in a shot Experimental results indicate the proposed method performs well in key-frame extraction with high efficiency关键词：key-frame;visual attention model;adaptive4218|255|0更新时间：2024-05-07
- 摘要：This paper presents a global alignment model based image mosaicing method for camera-captured document images, and it can be used to combine multiple overlapping document images into one large image It corrects the perspective distortion with the estimated vanishing points, and there exists only an affine transform between two adjacent images Then, it adjusts the distance of featurepoints to distribute them as evenly as possible in the overlapping regions Thirdly, it uses local alignment constraints of all the overlapping image pairs to construct global alignment model, thus, to eliminate the error accumulation In order to reduce alignment error of overlapping area, a binary weighted function is used to blend the overlapping region of image pairs This method is unique because it does not require the calibration of the internal/external camera parameters in advance and does not restricting the camera position, thus allowing greater flexibility than scanner-based or fixed-camera-based approaches. It can produce a high resolution and accurate full page mosaic from small image patches of a document关键词：document image mosaicing;image alignment;error accumulation;perspective distortion3653|291|0更新时间：2024-05-07
- 摘要：Considering the dissatisfactory precision and stability of primary M-Estimators, which depends entirely on the original matrix obtained by the method of least squares, an improved M-Estimators algorithm for estimating the fundamental matrix was studied The new algorithm obtained a more precise original matrix by calculating the distances between the matching points and the corresponding epipolar lines Then the mismatch and outliers in the original matching points set were eliminated through the precise original matrix and the nonlinear optimization for the new matching points set was carried out with Torr-M-Estimators Finally the accurate matching points set and the epipolar geometry can be gained Through a mass of experiments on simulated data and real images in the case of mismatching and Gaussian noise, the comparing results between the algorithm and other robust methods indicate the algorithm not only improves the estimating precision but also shows the good robustness关键词：fundamental matrix;robustness;precise original matrix;M-Estimators;LMeds(least median of squares)4091|219|0更新时间：2024-05-07
- 摘要：In recent years, technical bottleneck of traffic information collecting and transferring have basically been broken through, but large amounts of real-time traffic information cannot be well matched and fused with road network spatial database, which results in low efficiency in concerned applications, such as vehicle navigation system, public travel information service platform, logistics system and other ones in location based service（LBS）and intelligent transportation Systems（ITS). This paper aims to integrate real-time urban traffic information represented in Chinese natural-language with road networks spatial database, analysze linear reference methods embedded in multi-source and heterogeneous traffic information, as well as present an improved maximum matching algorithm to understand traffic information in Chinese natural-language and match it with spatial database automatically. Finally, a prototype system is developed to validate the approaches.关键词：traffic information fusion;natural language understanding;linear reference method;maximum matching algorithm3658|234|0更新时间：2024-05-07
- 摘要：Efficiency and accuracy are undoubtedly two exclusive indices for shortest path algorithms. The practical analysis on different algorithms solving the Kth shortest path problems has not invoked much research. The comparison on running efficiency between several theoretically rigorous algorithms and the amount of accuracy loss and efficiency improvement brought by lossy Kth shortest path algorithms have remained quantitatively under investigated, and the trade-off between them has not been analyzed in detail. In this paper, with a systematic classification of the most popular Kth shortest path algorithms, the author discussed the characteristics and time complexity of these algorithms, and analyzed and compared their efficiency and accuracy with a real roads network. The author argued that the theoretically rigorous Kth shortest path algorithms, generally lack practical value, only the multi-label setting algorithm can be applied for some applications requiring lossless accuracy. Some lossy Kth shortest path algorithms, greatly improve the efficiency with only little accuracy lost The bidirectional-search algorithm was argued worth paying more attention in practical applications.关键词：Kth shortest paths algorithms;roads network;efficiency;accuracy3895|246|0更新时间：2024-05-07
- 摘要：Ray-casting is a widely used basic volume rendering algorithm. It can get high quality image but suffers from high computational complexity. A single-pass ray-casting algorithm is developed based on GPU(graphics processing unit), and on the basis of this algorithm an accelerated algorithm is proposed based on space leaping technique. Noncontributing region in the volume data coded with octree data structure can be eliminated by space leaping. Then the burden of GPU may be reduced effectively. Only one fragment program is needed to generate ray direction, cast ray, skip empty voxels and terminate ray, et al. The experimental results illustrate the algorithm can significantly accelerate the reconstruction of the volume data containing a lot of empty voxels.关键词：GPU;space leaping;acceleration;ray-casting4484|236|0更新时间：2024-05-07
- 摘要：Medical image segmentation plays an important role in practical applications such as diseases diagnosis, surgical planning, and surgical guidance. In this article, we propose a fast medical image segmentation method based on the policy evolution level sets. Our evolution policy is to calculate the energy directly and check if the energy is decreased when we switch a point from the outer contour to the inner contour (or vice versa). By scan points of inner and outer contour, make the curve or surface move inward or outward to go to the boundary of object. This approach differs from the previous methods in that we do not need to solve PDEs, it can improves the computational speed dramatically. The problem (the local minimums and scan the whole image) of energy function calculate method is solved. At last some segmentation experiments is make on medical image in 2D image and 3D volume, and it demonstrated that our algorithm is fast and precision.关键词：level set;3D segmentation;policy evolution3288|234|0更新时间：2024-05-07
- 摘要：With the increasing requirements of video monitoring system’s intelligence, tracking moving object has been a hotspot. It is a problem needs be to addressed that how to detect object’s depth moving state（the size of object) only with 2D video signal and control the camera’s zoom. This paper proposes a method of tracking moving object in 3D-space based on genetic searching algorithm. The object’s 3D contour feature can be searched and matched using genetic algorithm to detect the location and the size of a moving object. Then, the camera’s zoom can be controlled. This method can ensure a relatively stable size of the moving object in the video screen and improve the quality of the monitoring system.关键词：template matching;genetic algorithm;3D contour feature3589|232|0更新时间：2024-05-07
- 摘要：Face recognition has aroused great concern for decades since it serves as a significant part in the fields of human-machine interaction as well as bioinformatics. Facial feature extraction is one of the key steps in face recognition system However, this step is characterized as being easily influenced by variations in face images such as illumination condition and expressions. In order to address this problem, a method that utilizes（2D)2PCA to extract facial features on the sub-bands obtained via wavelet packet decomposition（WPD) is proposed. There are three contributions:（1) take all multi-resolution sub-bands as research objects;（2) choose ‘successful’ sub-bands based on recognition rates;（3) propose a sub-band fusion method. Firstly, sub-bands are acquired by two-level WPD, then the feature matrixes of all sub-bands are calculated by（2D)2PCA, and further used to obtain recognition rates with the nearest neighborhood classifier. Thirdly, ‘successful’ sub-bands are chosen based on their recognition rates and fused to complete the task of face recognition. Finally, intra and extra experimental comparisons using samples of CMU PIE and Yale indicate that the proposed method gain satisfactory results and fusing sub-bands on multi-resolution can improve recognition performance.关键词：face recognition;multi-resolution;WPD;（2D)2PCA;feature matrix;sub-band3960|249|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