最新刊期
卷 16 ， 期 9 ， 2011
- 摘要：Imaging in the atmosphere is often degraded by scattering due to atmospheric particles such as haze，fog and mist.The task of visibility enhancement techniques is to correct for the loss of contrast and color fidelity，which can lead to large improvement in image quality.In this paper，we present a review on established approaches for visibility enhancement of images taken under foggy and hazy conditions.These approaches are grouped into two categories，namely，physics based and non physics based.We illustrate them in detail and then characterize their strengths and limitations，respectively.Additionally，some reference-free objective quality metrics are provided for quantifying the performance related to contrast and chromatic diversity.Finally，further discussions are pointed out on technical challenges and future development关键词：image defogging;visibility;color constancy;atmospheric scattering model;reference-free objective quality metrics7948|1043|0更新时间：2024-05-08
- 摘要：Aiming at solving the target fade-out and ghosting problems widely existed in the scene-based correction algorithms of infrared focal plane array (IRFPA)， a new adaptive correction algorithm based on motion guidance is proposed. Firstly， a background image and bad-cell template is prepared for the real-time correction. Next， a variable iterative step instead of a constant step in the steepest descent method is employed and adaptively adjusted according to the temporal variance. Finally， the gain correction coefficients are updated under the direction of variable iterative step and motion information， and then the image is corrected by the updated coefficients. The experimental results using real IRFPA infrared images show that the developed algorithm can not only effectively solve ghosting problems and inhibit the target fade out， but also eliminate the fixed-pattern noise well.关键词：infrared focal plane array;non-uniformity;adaptive correction algorithm;motion guidance3065|195|0更新时间：2024-05-08
- 摘要：Embedding capacity is one of the important factors that affect the information hiding perfomance.An improved information hiding scheme for VQ-compressed index map images based-on joint neighboring coding(JNC)technique is proposed to increase the embedding capacity.The proposed method embeds secret data by using the difference values ranging between the current VQ-compressed index and its four neighboring indices.Experimental results show that the proposed method increases the embedding capacity significantly compared with Chang’s and Yang’s method.The proposed method obtains 2.11 to 3.16 times embedding capacity compared to Chang’s method with only a 5.13%to 13.10%increas of the bit rate.As a result，the value of embedding-efficiency is increased on average from 0.10 in Chang’s method to 0.26 in the proposed method.关键词：information hiding;vector quantization;joint neighboring coding;embedding capacity;bit rate;embedding-efficiency3142|187|0更新时间：2024-05-08
- 摘要：The encoding process of vector quantization (VQ) is computational expensive. To reduce the distance computations between vectors， many methods use inequality relations to make distance estimations with less computational cost. This paper aims at extending the method using two-bounds triangle inequality proposed by Chang C C et al by improving reference the vector selection. The experimental results show that our proposed method can increase the ratio of codeword rejection by 3.735%9.976%， and reduce the encoding time by 6.03%35.25%.关键词：vector quantization;fast codeword search;reference vector;encoding performance2507|182|0更新时间：2024-05-08
-
Research on vehicle shadow segmentation with object knowledge constraint based on multi-color spaces
摘要：Intelligent Transportation System research aims at managing traffic flow by providing vehicle information in the scene.Moreover，moving vehicle detection in a video sequence is a very difficult question because of vehicle shadows.In this paper，first，a novel background model update is proposed，which uses object knowledge which is learnt by previous frame to renew background model.Second，a new color space is proposed for vehicle shadow segmentation，which is VIRGBC color space.The proposed technique exploits spectral and geometrical properties of shadows，by fusing all the shadow sensitive features extracted from different color spaces to obtain the VIRGBC color feature.Finally，we use conditional random model to train each channel mask of the VIRGBC color space，and fuse every channel result to obtain the segmentation result by harmonious mean.Experimental results demonstrate the effectiveness of the proposed method，and particularly interesting are the results obtained by the new approach in terms of area under precision recall curve (93.5%)，which is better than those obtained by other state-of-the-art method.关键词：shadow segmentation;color space fusion;conditional random field;traffic surveillance2689|273|0更新时间：2024-05-08 - 摘要：A conditional random field (CRF) model is used to incorporate different feature potentials of objects for multi-class object recognition and segmentation in natural images.By using an over-segmentation algorithm，we propose a new region based CRF model called R-CRF model.We train our model on annotated samples by using Joint-boost algorithm and investigate the performance of the theme based R-CRF model for class based pixel-wise segmentation of images.We compare our results with recent published results on the MSRC 21-class database.The result shows that our theme based R-CRF model significantly outperforms the current state-of-the-art.Especially，by introducing theme and regions，our model obtains greatly improved accuracy of structured classes with high visual variability and fewer training examples.关键词：object recognition;image segmentation;multi-class;CRF model;joint-boost3086|271|0更新时间：2024-05-08
- 摘要：In mining and analyzing high-dimensional data task， when only a small number of pairwise constraints including must-link and cannot-link are available， supervised dimensionality reduction methods tend to perform poorly due to the lack of data labels. In such cases， unlabeled samples could be useful in improving the performance. In this paper， we propose a novel semi-supervised locality dimensionality reduction algorithm (SLDR) in terms of pairwise constraints and abundant unlabeled samples. Specifically， SLDR can effectively use local information of the data and pairwise constraints to find a projection. After the data is projected onto a low-dimensional space， instances involved by cannot-link constraints are far apart， while instances involved by must-link constraints are close to each other. Moreover， the intrinsic geometric information of the data is preserved. In addition， SLDR can be extended to nonlinear dimensionality reduction scenarios by the kernel trick， which is applied to reduce the dimensions of highly nonlinear data.关键词：pairwise constraint;locality information;dimensionality reduction;discriminant analysis2468|204|0更新时间：2024-05-08
- 摘要：Existing methods based on local features cannot recognize objects in real-time while keeping a high recognition rate.Considering that many local features are unstable，unreliable，or irrelevant，we are able to select a small subset of features used for recognition by correctly matching features in training images.A new，robust，and stable method based on a bag-of-features is proposed in this paper.Distinctive features are selected by an unsupervised preprocessing step.Our experiments demonstrate that this selection approach can reduce the amount of local features and reduce the memory requirements，while allowing an average of 4%of the original features per image to provide matching performance that is as accurate as the full set.The method can meet real-time requirements since the time required for matching has been reduced from seconds to tens of milliseconds.关键词：local features;bag-of-features;object recognition;real-time3128|690|0更新时间：2024-05-08
- 摘要：Aiming to reduce the limitation in local background covariance matrix estimation of RX algorithm，an improved RX (I-RX) algorithm is proposed for anomaly detection in hyperspectral imagery.Based on a singular value decomposition (SVD)，We firstly project the hyperspectral imagery onto the background orthogonal subspace to obtain the remaining imagery which only consists of noisy background and anomaly.On this basis，by calculating the spatial rank depth value of every sample，the remaining imagery can be divided into two sample sets：noise background set and potential anomaly set.Using the noise background set to estimate the background covariance matrix of the whole imagery and the potential anomaly set as test examples to be detected whether has anomaly or not.Numerical experiments are performed on simulated data and real hyperspectral data.The ROC curves demonstrate that the detection probability of I-RX algorithm is about 2 times than RX algorithm at the same false alarm rates.关键词：anomaly detection;singular value decomposition (SVD);background orthogonal subspace;spatial rank depth;hyperspectral imagery3754|217|0更新时间：2024-05-08
- 摘要：In the scene matching aided navigation system based on the modified Hausdorff distance，a roughness-to-precision hierarchical matching using multi-resolution wavelet decomposition technique is often used to achieve a real-time matching.However，this method is not suitable for the case of small image.Based on the characteristic of Hausdorff distance，a pixel-jump searching method is proposed in this paper.Furthermore，by comparing the pixel-jump searching with wavelet decomposition，an adaptive matching method is brought forward.Simulation results show that the proposed fast adaptive matching method takes less time than multi-resolution wavelet decomposition matching，and that the match point is correct even when the image dimension is small.2895|249|0更新时间：2024-05-08
- 摘要：Many factors in outdoor environment，such as light，shadows and trees blocks etc.will affect the video unusual event detection based on the hidden Markov model(HMM).So using the independent component analysis (ICA) and HP (Hodrick-Prescott) filter，a new HMM video unusual event detection method is proposed.This method first employed ICA to construct a normal video feature subspace，and projected the image sequence into this subspace to achieve data reduction.Then the trend component in feature sequence caused by the environment factors is cancelled by the HP filter.So to overcome the adverse environmental factors，the effect of the proposed method is improved.Using the video data in a forbidden road for motor vehicles，the test result shows the proposed method can detect the video unusual events effectively in the complex outdoor environments.关键词：unusual events detection;independent component analysis (ICA);HP filter;hidden Markov model (HMM)3007|207|0更新时间：2024-05-08
- 摘要：As the feature dimension increases，the original PMK suffers from distortion factors that increase linearly with the feature dimension.This paper proposes a new method by consistently dividing the feature space into two subspaces while generating several levels.In each subspace of the level，the original pyramid matching is used.Then a weighted sum of every subspace at each level is made.To optimize the added kernel matrix，we get a p.s.d.kernel which can be used in kernel based learning methods (such as SVM).Experiments on dataset Caltech-101 and ETH-80 show that：compared with other related algorithms which need hundreds of times of original computational time，It takes only about 46 times of original computational time to obtain the same accuracy by using the method of DP-PMK.关键词：dimension partition;bags of features;SVM;pyramid matching;Kernel function;object recognition3267|209|0更新时间：2024-05-08
- 摘要：In this paper，based on biologic visual information processing，an improved Itti’s visual attention model is proposed for image retrieval.By considering the relation between texture and visual perception，a coarseness map is constructed as a primary visual feature of the new visual attention model.Therefore，by the improved Itti’s model，we can obtain 50 feature maps，and extract the distributions information over the local binary pattern histogram fourier (LBP-HF) of each feature maps to get the high dimensional features.Finally，the locality preserving projections (LPP) is utilized for dimensionality reduction，and the low dimensional feature having both the local geometry and the discriminate information is used for image retrieval.Experimental results show that the proposed approach has good retrieval performance.关键词：image retrieval;visual attention model;visual feature map;LBP-HF operator3419|234|0更新时间：2024-05-08
- 摘要：Dense matching can be generalized as a process of energy minimization.The computation of matching cost is an independent step for Dense matching.We analyzed a recent well-performed weighted local matching algorithm，with which we proposed an improvement of core steps named the computation of geodestic distance.After introducing the advancements of the new algorithm in detail，we show that：nearly 1/2 time cost of matching cost computation was saved without loss of precision in our experiment.关键词：geodestic distance;weighted window;matching cost;middlebury2787|206|0更新时间：2024-05-08
- 摘要：An optimal viewpoint selection method for volume rendering based on shuffled frog leaping algorithm is presented.Utilizing the opacity，luminance and structure features of the projected views of three-dimensional volume dates，a viewpoint evaluation function is constructed to identify the importance of voxels and the structural information within the volume data.Combined with this function，shuffled frog leaping algorithm is used to optimize the computation process of the optimal viewpoint selection，so as to acquire a global optimal viewpoint or a series of optimal viewpoints.The experimental results show this method can not only focus on important structural information and region of interest more effectively，but also improve the convergence rate and accuracy，while having a standout performance to select a global optimal viewpoint.Thus it can be used to guide the volume rendering for large scale volume datasets.关键词：volume rendering;viewpoint selection;viewpoint evaluation;shuffled frog leaping algorithm;structural information;entropy2837|197|0更新时间：2024-05-08
- 摘要：Space block strategy is the effective method in finding K-nearest neighbors.However，during dividing the min-max box of the dataset，the size of sub-cubes mainly is decided by the K value，and the min-max box is needed to divide again along with the change of K value in existing algorithms，it has affected time efficiency and stability of algorithm.Using the advantages of space block strategy，a new algorithm for finding K-nearest neighbors is presented with establishing the space index of scattered points as the spatial division goal.The min-max box of the dataset is divided into a set of uniform sub-cubes in advanced，the structure of sub-cubes is only related to the scattered points and the default parameter，the same points cloud data just is distributed one time.In the course of searching，the dynamic sphere is builded using the test point as the center of sphere，and judging eligible sub-cubes in order to find K-nearest neighbors.The experimental results show that the new algorithm is more superiority than existing algorithms in points cloud distribution and searching time efficiency，the stability of searching time of random points，and the adaptability of different K value.关键词：spatial sub-cubes;K-nearest neighbors;dynamic sphere;comparison of algorithms2663|202|0更新时间：2024-05-08
- 摘要：Facing the sparse LiDAR (light detecting and ranging) data of the New Orleans area in America，a new residential building model reconstruction method is proposed.The main contributions of this work are the automatic isolation of roof points and the roof type recognition.Using the minimum bounding contour to describe the outline of residential buildings，we are able to automatically identify individual buildings from clustered residential areas.Then，based on the relationship of normal vectors，building types are recognized； through model matching，the buildings models are reconstructed.Experiments show that our method can successfully reconstruct residential buildings given relatively sparse LiDAR samples in a reasonably short time.关键词：building reconstruction;minimum bounding contour;LiDAR;roof isolation;model matching3064|190|0更新时间：2024-05-08
- 摘要：During the course of high-resolution remotely sensed images information recognition for agriculture and forestry，image segmentation is an important procedure since its results decide directly the effects of land-parcel auto recognition.After analyzing and comparing the segmentation effects of general methods，this paper draws the conclusion that mean shift algorithm can be applied to agriculture and forestry application and has been applied to multi-scale image segmentation because it can extract exactly the edges of the land-parcel.By analyzing the principle of the mean shift image segmentation，this paper improves the algorithm by using two methods，one is to implement the multi-scale merge of the segmentation parcels，and the other is to implement parallel segmentation in message passing interface (MPI) and OpenMP (OMP) model environments.And the applications in agriculture and forestry show that the precision and speed can meet the need of Agroforestry.关键词：agriculture and forestry land-parcel;multi-scale segmentation;mean shift;MPI;OMP2583|274|0更新时间：2024-05-08
- 摘要：The cloud is a natural phenomenon.It widely， frequently， irregularly appears in the earth sky and thus in the satellite imagery.The cloud in the remote sensing images largely reduces remote sensing image quality and usage ratio.The presence of clouds in satellite images is a significant obstacle to land surface studies.Clouds in satellite images must be precisely identified prior to any further analysis in any case.Cloud detection is a necessary step in the image data analysis processing.Therefore accurately extracting the location of the cloud area becomes a very meaningful research subject.This paper has proposed a cloud detection algorithm for MODIS image based on HSV color space.The HSV color space is first introduced in cloud detection study.Based on the different objects spectrum， MODIS instrument channel characteristic and the HSV positive transformable formula， the algorithm firstly has processed the HSV positive transformation for the pseudo-color image composed of the band 1， band 6 and band 26.Then to carry a threshold for the hue， we can get the final cloud detected results.The algorithm is simple， impersonal and fast.The results show that the algorithm is applicable to all the different scenes and seasons and has basically completed.关键词：MODIS;cloud detection;RGB;HSV;color space3284|219|0更新时间：2024-05-08
- 摘要：A dynamic geological modeling method based on topology reasoning among sections is presented in this paper in order to resolve the geological section correspondence problems，which results in the bottlenecks of automatic modeling，and enables the 3D stratified geological solid model to be updated dynamically.With topology relationship construction and reasoning on the three-dimensional surface，this method increased the dimensions of the geological section correspondence，and implemented dynamic three-dimensional geological modeling.This method had been used in Nanjing urban geologic three-dimensional modeling.The result indicated that it was suitable for many kinds of section data，and enhanced the three-dimensional modeling efficiency.关键词：geological section;3D geological solid;topology reasoning;dynamic modeling2693|214|0更新时间：2024-05-08
- 摘要：The paper presents an interpolation method for surface air temperature (SAT) based on data fusion of multiple sources.It should check whether there is a significant correlation between primary and secondary variables firstly.Three multivariate geostatistical algorithums which includes collocated cokriging (CCK)，simple kriging with varying local means (SKlm) and kriging with an external drift (KED) were introduced to incorporating ancillary information into the spatial prediction of SAT.The method was illustrated using monthly mean temperature data from more than 720 meteorological stations in China in August 2008，and cross validation was performed to evaluate the performance of the map prediction quality.The results show that:Accounting for both land surface temperature (LST) from remote sensing and digital elevation model (DEM)，used as ancillary spatial information in three algorithms，outperforms accounting for only one ancillary data.Among all different methods，SKlm and KED incorporating LST and DEM have produced the best results，this is because:(1) LST is better to indicator the local trend of SAT.(2) DEM prefers to indicator the global trend of SAT.(3) Both SKlm and KED considering SAT with a non-tationary spatial distribution have better performance than others.关键词：surface air temperature interpolation;information fusion based on multiple sources data;multivariate geostatistics3634|205|0更新时间：2024-05-08
- 摘要：Commission errors often arise when Gaussian mixture models (GMMs) are applied to detect moving objects in situations where foreground and background have similar colors.To improve the robustness of the segmentation method，a moving object detection method is proposed by combining a Gaussian mixture model with the wavelet transform.The wavelet transform is employed to extract texture information and a GMM is employed to update the background.Color information and texture information are integrated for segmentation.The method improves the abilities of convergence and stability，and also decreases commission errors occuring in methods which only use color information.Experimental results indicate that the proposed method is superior to the traditional GMM.关键词：Gaussian mixture model;wavelet transform;object detection4617|283|0更新时间：2024-05-08
- 摘要：Due to dissatisfactory results of segmenting noisy images，a new variational method for segmenting images corrupted by various noises is presented.First，the energy function of the proposed model based on Chan-Vese(CV) model is modified.Then in order to improve the computing efficiency and the accuracy of the proposed energy minimization problem，a new auxiliary variable is introduced，which is combined with the level set method.By applying the multiplication operator splitting (MOS) numerical algorithm，good results are obtained.Finally，compared to the CV model，our results show that the proposed method hurdles the influence of noises，which is effective for segmenting the noisy images.Less iterations are needed，and our approach is faster than the CV model.Furthemore our model has higher object segmentation accuracy.关键词：noisy images;image segmentation;auxiliary variable;MOS algorithm4128|203|0更新时间：2024-05-08
- 摘要：The difficulty of counting the number of people using video surveillance lies in the correct segmentation of each person in crowded situations.For this problem，we train a classifier to detect head candidates so as to distinguish between people walking closely under the overhead mounted camera.The static false detected candidates are removed according to the motion feature of the head，while the dynamic ones are eliminated due to their low detected frequency and discontinuous response positions.Finally，an easy and effective crossing tracking scheme is proposed for counting.We conducted experiments in various complex situations.It shows that the counting rate is over 95%.Compared to the traditional method，the proposed means can segment and count the people more accurately in cases where multiple people are crossing or crossing with items.It is therefore much more suitable for people counting in real life situations.关键词：AdaBoost;head recognition;inter-frame differencing;detection response features;object tracking4742|282|0更新时间：2024-05-08
- 摘要：An alternating minimization method for the reconstruction of MR images with wavelet sparsity，including the convergence analysis of this algorithm，is presented in this paper.Our algorithm is base on a half-quadratic penalty method.The optimization problem with wavelet sparsity is splited into two sub-problems:the X-subproblem and the Y-subproblem.The solution for the original problem can be obtained by alternately solving the two subproblems，in which the Y-subproblem is solved via a 1D soft-thresholding or shrinkage and the X-subproblem is solved by a Fourier transform approach.Thus，the splitting algorithm for solving the original problem is generated.The phantom image and some real MR images are employed to test our approach in the numerical experiments.The alternating minimization algorithm is also compared to the state-of-the-art algorithm，operator splitting algorithm.The experimental results demonstrate that the alternating minimization algorithm has not only a greater signal-to-noise ratio(SNR)，but has also less relative errors and is faster than the operator splitting algorithm.Therefore the alternating minimization method is a fast reconstruction method for sparse MR images.关键词：compressed sensing;optimization;sparse image;image reconstruction4918|225|0更新时间：2024-05-08
- 摘要：Two-dimensional animation is a labor-intensive process， where automatic keyframe in terpolation (inbetweening) is a research focus.Two-dimensional shape-blending plays a key role in keyframe inbetweening. Most of the existing 2D shape-blending algorithms are designed for a single closed polygon. When they are applied to complex shapes containing multiple regions， overlaps between different regions may occur. Our algorithm first obtains closed regions of input keyframes and calculates the matching relations between these regions. Then we iteratively construct a local coordinate system relative to its parent region for each child region. Then the hierarchical constraint relation can be used to get rid of “motion” mismatches between parent-child areas. Experimental results show that our algorithm can synthesize naturally smooth inbetweens and is efficient enough for practical demands.关键词：in-betweening;shape blending;region matching;hierarchical constraints4757|192|0更新时间：2024-05-08
- 摘要：In this paper we proposes a new method for comparing the quality of defogged images based on the halo artifacts，the color shift，and other issues which often appear during the defogging process.the canny edge detector is used to denote the detail performance of defogged images，is utilized to denote the color performance of defogged images，the is used.We use the structural information of the reflection image to indicate the structural performance of the defogged images.Using all these three factors，the effect of defogging methods can be evaluated objectively.Experimental results show that the proposed method，which is based on the visual differences of images before and after defogging，can compare the effect of different defogging methods and is consistent with the subjective perception.关键词：defogging;image quality assessment;valid edge intensity;color recovery5712|339|0更新时间：2024-05-08
- 摘要：Stereo image quality assessment is the basis of perception and is also the foundation of stereo video system design. Since human beings are the final receiver of the images， it is essential that the image quality assessment matches the characteristics of the human visual system (HVS). An algorithm for stereo image quality assessment is proposed based on the basic HVS characteristics such as visual non-linearity， contrast sensitivity， multi-channel model， masking effect， and so on. A five level wavelet decomposition is first performed on the images. The spatial frequency of an image is divided into six bands to vary the spatial frequency of the initial image according to the characteristic of the masking effect. Then the similarity measure for each of the frequency bands is evaluated. According to the contrast sensitivity characteristics， the quality evaluation results of each band are weighted to get the final quality assessment scale. The experimental results show that the proposed algorithm outperforms conventional objective image quality methods， and has good agreement between objective and subjective perception. It may reflect the image quality and stereo sense.关键词：image quality assessment;HVS;wavelet transform;stereo sense5185|207|0更新时间：2024-05-08
- 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