最新刊期
卷 18 ， 期 2 ， 2013
- 摘要：Population aging is a current social factor which impacts significantly on the future of most countries. To relieve such economic and social pressures, one feasible way is to automatically understand the activities of daily living (ADL) of eldly people by using the highly developed information and communication techniques, so as to provide living assistance for eldly people to prolong their independent living time in home. ADL happens under unstructured natural environment. ADL recognition and understanding is actually a process of observation, analysis, inference and decision in space-time domain under dynamic environment. It essentially requires that the computer system should have the human-like cognitive competence. Therefore, ADL recognition and understanding is a confluence of interdisciplinary research (such as computer vision, cognitive computing, and pervasive computing) and real-world applications to cope with the problem of population aging. The goal of this paper is to discuss the challenges and basic scientific problems in ADL recognition and understanding, and to discuss possible future directions to reduce the gap between the start-of-art and requirements of ADL recognition in real world.关键词：population aging;ambient assisted living;activity of daily living (ADL);ADL recognition and understanding3573|335|0更新时间：2024-05-07
- 摘要：The common goals of human vision and computer vision are to understand the world via visual information. Effective utilization of all spatial-temporal information and understanding the behaviors of interesting objects in a scene is a current focus in computer vision research. In this paper, we provide a general introduction for the basic situation, main concepts, research issues, typical techniques and developments of this area, hope to draw the attention of research community, to jointly work in it and to push forward the front of computer vision.关键词：spatial-temporal techniques;action primitives;action;activity;event;behavior;image understanding4423|376|0更新时间：2024-05-07
- 摘要：It has been a promising methodology that designs machine vision algorithms based on the vision mechanism of primate. In this paper, from the intersection points of vision mechanism and machine vision algorithms, we summarize two categories of important vision mechanisms or models, and present their corresponding machine vision algorithms. 1)Cooperative learning and competitive learning: machine vision algorithms motivated by the models typically include stereo vision, neural networks and sparse coding. 2)Simple cell and complex cell: machine vision algorithms corresponding to the models focus on HMAX feature, SFIT feature and deep belief networks.关键词：vision mechanism of primate;machine vision algorithm;cooperative learning and competitive learning;simple cell and complex cell3346|388|0更新时间：2024-05-07
- 摘要：Vision-based human action recognition is currently one of the most active research fields. It has many promising applications such as intelligent surveillance, perceptual interface and content-based video retrieval. Even though, some hurdles still slower the development of action recognition, which the actions are often observed from arbitrary camera viewpoints in realistic scene. So view-invariance is important for action recognition, growing number of research groups to pay more attentions to research related to the view-invariant issue. This paper provides a survey on view-invariant recognition of poses and actions. The improvements of this topic in the last several years are discussed in detail from four aspects: space-time based methods, state-space approaches, dimension reduction based methods, and trajectories based methods. Public available standard datasets are presented, and the concluding discussion assesses the progress so far, and outlines some future directions.3734|370|0更新时间：2024-05-07
- 摘要：In the process of image compression, the discrete cosine transform (DCT)makes the image blurry. With the frequency warping, the warped discrete cosine transform (WDCT)overcomes this shortcoming. However, it is difficult for traditional algorithms to select the frequency warping parameters. Therefore, in this paper, we present a new WDCT image compression algorithm based on an evolutionary algorithm (EAWDCT)to select the parameters with the global optimization of the evolutionary algorithm. To improve the convergence speed and to gain the best parameters quickly and accurately, a heuristic method is used to design the chromosome coding methods and evolutionary operators according to the parameters characteristics. The experiments show that the new algorithm (EAWDCT)improves the peak signal to noise ratio and signal to noise ratio while reducing the root-mean-square error effectively.关键词：image compression;warped discrete cosine transform;frequency warping;evolutionary algorithm2908|253|0更新时间：2024-05-07
- 摘要：The results of image edge detection directly affects the subsequent image processing operations. Effectively extracting the information carried by the image target is the main purpose of the image edge detection. In order to realize the extraction of a target outline, an image edge detection method based on a Euclidean distance graph is proposed. The Euclidean distance between points in the image is calculated, and the distance graph is obtained. The scenery outside the outline is well described in the graph. The scenery outline can be effectively extracted by using an improved Canny operator for edge detection. Compared with other general edge detection algorithms, the method can effectively realize the detection of a target's outline. Because small and trivial details are suppressed, the method gives a better basis for further image target outline recognition.关键词：Euclidean distance;distance graph;improved Canny operator;edge detection5010|589|0更新时间：2024-05-07
- 摘要：In order to eliminate fog or haze in outdoor images in real time, a fast haze-removal algorithm is proposed based on the theory of dark channel prior. By improving the transmission repairing method, the computational effort is reduced. Furthermore, by self-adaptively controlling the values of transmission-rate lower limits and the global atmospheric lights, this algorithm could be used in more image dehazing applications. Finally, a post-processing procedure is adopted to enhance the visual effect of the dehazed image. Experiments show that the presented method processes the fog image effectively and fast,making the algorithm easy to achieve real-time capabilities.关键词：haze-removal;dark channel prior;transmission;imaging enhancement5650|462|0更新时间：2024-05-07
- 摘要：Choosing the right dictionary used for sparse coding has an important effect on image reconstruction and pattern classification. Therefore, a new sparse representation algorithm based on Gabor Feature Set Discriminative Dictionary Learning is proposed for image recognition. Considering that Gabor feature is robust to variations of illumination, expression, and pose, the proposed method first extracts the image Gabor features with multi-scale and multi-orientation.Then it uses the augmented Gabor local feature matrix whose dimension has been reduced to construct the initial feature dictionary. This reduction is based on the Fisher discrimination criterion. A structural dictionary, whose atoms correspond to the class labels, is learned so that each sub-dictionary of the learned new dictionary is a good representation of the samples from the corresponding class. Meanwhile, the Fisher discrimination criterion is imposed on the coding coefficients so that they have small within-class scatter but big between-class scatter. Consequently, a new classification scheme associated with the proposed method is then presented by using, the discriminative information and sparse coding coefficients. Experiments on three types of databases show that the proposed method is valid and efficient.关键词：sparse representation;sparse pattern classification(SPC);gabor feature;Fisher dictionary learning4318|319|0更新时间：2024-05-07
- 摘要：The category information of the -nearest neighbor labeled samples is used, but the contribution of the test samples is omitted in the weighted -nearest neighbor method, which often lead to misclassifications. Aimed at the problem, a semi-supervised -nearest neighbor method is proposed in this paper. The method can classify sequential samples and non-sequential samples better than the -nearest neighbor method. In the decision process of classification, the information of c-nearest neighbor samples in the test set is used. So, classification accuracy is improved. The recognition accuracy of the method is 5.95% higher for sequential images in Cohn-Kanade face database, and 7.89% higher for non-sequential images in Cohn-Kanade face database than it of weighted -nearest neighbor method. The experiment shows that the method performs fast and has high classification accuracy.3332|293|0更新时间：2024-05-07
- 摘要：A minimum-distance discriminant projection (MDP)algorithm is proposed to address face recognition problem. Different from the classical linear discriminant analysis (LDA), the MDP is a manifold learning based dimensionality reduction algorithm. MDP first defines the intra-class similarity, weight, and the inter-class weight of each sample. The former one can measure the distance between each data point and the intra-class center, while the latter one does not only characterize the distance between the data point and the inter-class center but also can reflect the relation between the between-class distance and the within-class distance. Then, the high-dimensional data is mapped into a low-dimension space such that the points to within-class center distances are minimized while the points to between-class center distances are maximized simultaneously. At last, experiments on the ORL, FERET, and AR face databases show that the proposed algorithm can outperform other algorithms.关键词：face recognition;dimensionality reduction;linear discriminant analysis;locality preserving projections;minimum distance3627|338|0更新时间：2024-05-07
- 摘要：Most steganalysis methods for image LSB matching suffer from a drop in the performance when applied to different image databases, especially for JPEG-compressed and uncompressed covers, which is also the main restriction to prevent steganalysis from practical usage. In this paper,we specify, the JPEG compression of a bitmap by taking advantage of the quantized DCT alternating coefficients,which no longer obey the Benford's law very well. Then variations of steganalysis methods and parameter sets are exploited according to the number of channels and whether the image has been compressed before. The algorithms' performances get a great boost by the proposed working mode.关键词：information hiding;steganalysis;LSB matching;Pre-classification;Benford’s Law2897|248|0更新时间：2024-05-07
- 摘要：In order to overcome the weakness of tradition CS-LBP descriptor in rotation-invariant, in this paper, we propose a new CS-LBP texture spectrum descriptor from the perspective of the human visual. It integrates the rotation of texture pattern with the shift operation of string. Based on the concept of texture pattern equivalent, a more effective method is designed to enhance the anti-rotation ability of the new descriptor. Experimental results show that the new descriptor and method have with more rotation robustness, and improve the clustering accuracy of rotary texture image by 9% to 38%.关键词：center-symmetric local binary patterns;rotation invariant;texture spectrum descriptor;image feature;image recognition;feature extraction3424|250|0更新时间：2024-05-07
- 摘要：Pencil sketching is an art, which represents objects by single color lines. The artists represent light and shading, while depicting shape features through observing the structure of the objects. In this paper, we present a non-photorealistic rendering technique that automatically generates a pencil sketch painting with shape features. We first use the bilateral filter to remove the noise. The vector field is calculated from the luminance image. We then develop a method for constructing a smooth vector field, which preserves shape features of the original image. Finally, line integral convolution and enhancing the contrast of light and shadow are used to generate the texture of pencil sketch. The experimental results show the effectiveness of this method in depicting an illusion of depth and producing the contrast of light and shadow.关键词：non-photorealistic rendering;pencil sketch;shape feature;vector field;edge tangent flow;line integral convolution3749|338|0更新时间：2024-05-07
- 摘要：A common point of partial differential equation and graph theory based image segmentation methods lies in creating and optimizing their energy functions. From the viewpoint of creating energy models, statistical image features from nonparametric estimation are measured with Bhattacharyya metrics, which is further embedded into energy function construction in Geodesic Active Contour (GAC)and Graph Cuts (GC)models in this paper. The improved GAC and GC models benefit from the energy function based on the aforementioned metric, which introduces a pull-back strength into the GAC to prevent boundary leaking and to help the GC model in accurately estimating the distribution from small samples and unstable distribution function as well as extracting objects in more detail. Then, the proposed methods are applied to the medical image segmentation scenario and a bone and meniscus segmentation framework on knee MRI sequence is presented. In the experimental section, quantitative and qualitative comparisons are conducted respectively. Experimental results show the increased precision of our method in segmenting medical images such as knee MRI sequences, which are affected by the noise and the partial volume effect.关键词：active contour model;graph cuts;Bhattacharyya distance;nonparametric estimation;knee image segmentation3453|333|0更新时间：2024-05-07
- 摘要：With the development of remote sensing technology and imaging spectrometer, hyperspectral remote sensing images are widely used. However, the features of hyperspectral images have brought great difficulties for its classification and identification. One important research question is "How to select a group of bands from hundreds of bands of hyperspectral images, which are good for classification and identification?" In view of the above question, the existing band selection methods are analyzed, and a new method of hyperspectral imagery band selection is proposed, which is combined with genetic algorithm and ant colony algorithm. In the algorithm, the genetic algorithm is used to search for some better solutions quickly which initialize the information list of the ant colony algorithm. Then, the ant colony algorithm can effectively search for the best solution. In the part of the genetic algorithm, quaternary encoding is used, which makes encoding/decoding and genetic operation simple and uses less memory. In the part of the ant colony algorithm, subspace division is used to deal with hyperspectral images, reducing the search range of the ants. Which improves the search efficiency, and reduces the correlation and redundancy of the output band of hyperspectral image. The algorithm makes good use of the advantages of both genetic algorithm and ant colony algorithm and overcomes their defects, by consuming less time and outperfoming restraining method for band selection. An AVIRIS image was used for experiment with the proposed algorithm, which proves that this algorithm of hyperspectal dimension reduction is effective in terms of band selection performance and execution time consumption.关键词：hyperspectral imagery;band selection;genetic algorithm;ant colony algorithm4175|489|0更新时间：2024-05-07
- 摘要：Restoring the complete shadow of targets by inpainting techniques is helpful for target recognition in high-resolution SAR image. In the process of shadow inpainting for high-resolution SAR images, the traditional exemplar matching leads to the misjudgment of shadow regions and the inhomogeneity of transition regions. In this paper, we propose an adaptive target shadow inpainting method based on exemplar's similarity. The problem of shadow region misjudgment is solved by introducing inpainting region prejudgment according to the relative relationship between target and its shadow. Moreover, the adaptive inpainting method based on exemplar's similarity overcomes the problem of transition region's inhomogeneity. The experimental results show that the proposed method can effectively restore the missing target shadows in SAR images.关键词：SAR image;image inpainting;radar shadow;similarity;exemplar2965|251|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