92 results about "Shape context" patented technology

The invention belongs to the technical field of the pattern identification and image processing, and relates to a license plate detection and identification method based on the maximum stable extremal region, which comprises the following steps: extracting the maximum stable extremal region (MSER) to obtain text regions of candidate license plates; adopting one effective feature for description of each extremal region, classifying the extremal regions into 'text' regions and 'non-text' regions with a classifier obtained in prior training, and extracting a license plate from original images based on the characteristics of the structure of the license plate; and describing with the characteristics of the shape context, and completing the character identification through template matching. Since the maximum stable extremal region has the affine invariance, high stability and multi-scale features, and the regions are determined only according to the gray value and are not light-sensitive,the license plate detection and identification method with the maximum stable extremal region as the base is applicable to the complex background, and has the advantages of good stability and high identification rate.

A system, method, and computer program product for estimating human body pose are described. According to one aspect, a human figure silhouette is segmented from a depth image of a human actor. Contour points are sampled along the human figure silhouette. Inner Distance Shape Context (IDSC) descriptors of the sample contour points are determined and compared to IDSC descriptors of the feature points in an IDSC gallery for similarity. For each of the feature points, the sample contour point with the IDSC descriptor that is most similar to an IDSC of the feature point is identified as that feature point in the depth image. An estimated pose of a human model is estimated based on the detected feature points and kinematic constraints of the human model.

The invention relates to the technical field of navicular identification, in particular to a method and system for navicular identification based on interest target regions and SUSAN (Small Univalue Segment Assimilating Nucleus) edge detection. The navicular identification method comprises the following steps: S1, converting an image to be identified into a gray level image; S2, extracting an interest target region from the gray level image; S3, utilizing a SUSAN operator to perform edge detection on the interest target region to obtain edge information, and performing target contour tracing on the edge information to form an shape of the target to be identified; S4, obtaining shape context of the shape of the target to be identified; S5, matching the shape context of the shape of the target to be identified with shape context of the shape of a template, and determining whether the target to be identified is the navicular identification. The method solves the problems of identification of a navicular object on an ocean surface scene and difficulty in navicular identification for on-board cargo, and can be widely applied to monitoring of exploitation of marine resource, thereby satisfying the requirements of fishery administration.

A new measure to compare histogram-based descriptors, a diffusion distance, is disclosed. The difference between two histograms is defined to be a temperature field. The relationship between histogram similarity and diffusion process is discussed and it is shown how the diffusion handles deformation as well as quantization effects. As a result, the diffusion distance is derived as the sum of dissimilarities over scales. Being a cross-bin histogram distance, the diffusion distance is robust to deformation, lighting change and noise in histogram-based local descriptors. In addition, it enjoys linear computational complexity which significantly improves previously proposed cross-bin distances with quadratic complexity or higher The proposed approach is tested on both shape recognition and interest point matching tasks using several multi-dimensional histogram-based descriptors including shape context, SIFT and spin images. In all experiments, the diffusion distance performs excellently in both accuracy and efficiency in comparison with other state-of-the-art distance measures. In particular, it performs as accurate as the Earth Mover's Distance with a much greater efficiency.

The invention discloses a method for retrieving a similarity shape. The method comprises the following steps: (1) extracting shape outlines of an input retrieving image and an image to be retrieved in a database; (2) describing all shapes (including input retrieving shapes and shapes to be retrieved in the database) by using internal distance shape context descriptors; (3) matching all shapes (including the input retrieving shapes and shapes to be retrieved in the database) pairwise by using a dynamic planning method; (4) calculating by using a context sensitive similarity measurement method to acquire a new similarity measurement sequence; and (5) acquiring a result of shape retrieving. In the method, dissimilarity (distance) among pairwise shapes is not used as direct basis for shape retrieving, but the dissimilarity (distance) among initial pairwise shapes is improved by integrating internal differences of shapes and utilizing structure information in a shape similarity space, so that the accuracy of shape retrieving is effectively promoted.

The invention relates to a kind of method to matching the shape of persons' faces based on Shape Context. At first, preprocessing with the image pyramid and diffusion-filtering technology; then abstracting the boundary-outline information; carrying on the logarithmic coordinate transformation of the abstracted boundary-outline information to get bar chart of the logarithmic coordinate; calculating the value of Cs to get all the phase match points; calculating the value of the degree of approximation to judge the shape matching. The invention is characterized in simplicity, accuracy, being economical, expandability and so on, so the person-face shape matching can be used for entrance-exit control, safe verification, safe-guard monitoring, searching criminals and so on.

The invention discloses a visual inspection algorithm for an outline. The algorithm comprises the following steps that: extracting a workpiece sub pixel edge, carrying out denoising and outline supplementation, carrying out outline matching, and carrying out deviation calculation. By use of the visual inspection high-accuracy algorithm for workpiece outline defects, firstly, the workpiece outlineis extracted, then, the outline matching is carried out on the basis of shape context features, accurate matching is carried out according to an iterative closet point algorithm, and finally, a matching deviation is calculated through a neighbourhood method. According to the comparison of a great quantity of experiments, the visual inspection method for the outline can accurately detect the outline defects of any shape of the workpiece, and detection accuracy can be 0.5 pixel. When the method is applied to industrial detection, the processing and the production of precise firmware can be intelligently detected at high accuracy, and production efficiency is greatly improved.

A system, method, and computer program product for recognizing hand postures are described. According to one aspect, a set of training images is provided with labels identifying hand states captured in the training images. Inner Distance Shape Context (IDSC) descriptors are determined for the hand regions in the training images, and fed into a Support Vector Machine (SVM) classifier to train it to classify hand shapes into posture classes. An IDSC descriptor is determined for a hand region in a testing image, and classified by the SVM classifier into one of the posture classes the SVM classifier was trained for. The hand posture captured in the testing image is recognized based on the classification.

The invention provides a goods classification method and a goods classification device both based on image features. The method includes: extracting shape context features and SIFT (scale invariant feature transform) features of image feature points; according to the shape context features and the SIFT features of the image feature points, matching similarity of the feature points among images to acquire matching results of the similarity of the feature points among the images; acquiring the similarity among the images according to the matching results of the similarity of the feature points among the images; according to the similarity among the images, classifying all the images by a classification method based on increment category trees to acquire classification results; outputting the classification results serving as goods classification results. By the goods classification method and the goods classification device, users can quickly browse goods according to styles and can be prevented from repeatedly browsing goods with the same styles, and users' shopping experience can be improved.

The invention discloses a heterologous image accurate matching method, which comprises the following steps: firstly, respectively matching an infrared grayscale image and a negative image thereof witha visible light grayscale image, namely two to-be-matched image groups; secondly, for each image of the to-be-matched image group, extracting key points with invariable scales, and calculating LPQ feature vectors of neighborhoods of the key points; thirdly, performing weighted fusion on the SIFT feature, the shape context feature based on the angular point and the LPQ feature of each image in theto-be-matched image group, and performing initial matching on the images through a nearest neighbor ratio method; fourthly, removing mismatching points; and finally, integrating the matching resultsof the infrared grayscale image and the negative image thereof with the visible light grayscale image into a final matching result. According to the method, the context descriptor is designed to obtain global information of the image, meanwhile, in order to further improve the matching performance, the LPQ descriptor is utilized to obtain image fuzzy invariant texture features, and finally, the method can obtain a heterologous image accurate matching result.

The invention belongs to the technical field of pattern recognition, image processing and computer vision, and particularly relates to a frequency domain shape description method for image matching, recognition and retrieval. According to the frequency domain shape description method for the image matching, recognition and retrieval, an image region is divided into a plurality of sub regions, each feature point is used as a reference point to count the number of other feature points falling upon each sub region to obtain a column diagram, a power spectrum of the column diagram is calculated to be used as a shape description, the similarity degree between the shape descriptors of two images is calculated to obtain the similarity degree of the images, and therefore the image matching, recognition and retrieval are achieved. Experiments show that the shape context spectrum signatures of the frequency domain shape description method for the image matching, recognition and retrieval has good rotation and stretching invariance properties, and good robustness on noise and deformation, and is low in calculation cost.

The invention belongs to the technical field of computer multimedia. In order to provide a simple and practical multi-view video interpolation and three-dimensional modeling method, the invention adopts the technical scheme that: a time-space jointed multi-view video interpolation and three-dimensional modeling method is provided, a plurality of camera arrays are grouped in a spacing manner, and the three-dimensional model of a scene at every moment is reconstructed. The time-space jointed multi-view video interpolation and three-dimensional modeling method specifically comprises the following steps of: (1) obtaining unacquired frames between two frames through interpolation; (2) obtaining an image of unacquired visual angles at this moment by using a model-assisted weighting method; (3) calculating and extracting key points; (4) describing the extracted key points by using shape context and solving by the Hungarian method; (5) obtaining final interpolated frames through solving the problem of Poisson editing and optimization; and (6) reconstructing and rendering the three-dimensional model of the scene. The time-space jointed multi-view video interpolation and three-dimensional modeling method is mainly applied to the design and the manufacture of antennae.

The invention discloses an image identification method and system. The image identification method includes the steps of: S1. obtaining from a picture to be identified an image contained by the picture; S2. segmenting the image into a plurality of slices, each slice containing a character; S3. according to pixel points of each slice, calculating shape context corresponding to each slice; and S4. searching according to a characteristic value of the shape context, and finding a sample closest to the character contained by the slice as an identification result. The image identification method provided by the invention can overcome the defect in the prior art that identification of special characters such as transformed characters or handwritten characters is inaccurate, accurately identifies special characters which are scaled, rotated, overlapped and distorted, reduces required sample files, reduces time required by identification, and improves identification efficiency.

A method and apparatus is disclosed whereby a point on an ear impression model to be labeled is selected and a shape context is determined for that point. This shape context is then compared to average shape contexts for different regions on a reference ear impression model, also referred to herein as an ear impression shape atlas. A cost function is used to determine the minimum cost between the shape context for the selected point and one of the average shape contexts. Once the minimized cost is determined, the region label corresponding to the average shape context having a minimized cost is assigned to that point. In this way, points on the surface of an ear impression are classified and labeled as being located in regions corresponding to the regions on the ear impression shape atlas.

The invention discloses an action recognition method and device based on the shape context. The method includes the steps of firstly, collecting multiple frames of continuous video images of human body actions; secondly, extracting the human body outline in any frame of image in the extracted video images, wherein the human body outline is composed of a plurality of outline points; thirdly, calculating the shape distance between all the outline points in the human body outline and outline points in key outlines contained in a key outline base, wherein the key outlines are standard human body actions; fourthly, sequentially selecting the key outlines, closest to all the outline points in the human body outline, in the key outline base for all the outline points in the human body outline, and voting for the key outlines; fifthly, counting the number of votes for the key outlines in the key outline base, and recognizing the human body action in the video image as the standard human body action corresponding to the key outline with the largest number of votes. According to the method and the device, accuracy of recognition of he human body action can be improved.

The invention provides a workpiece image sparse stereo matching method based on an improved-type shape context. The method integrates a shape context capable of reflecting a point position distribution relationship and gradient direction histogram features capable of reflecting point gradient attributes, and mainly comprises steps: pretreatment, such as gray normalization and Otsu binaryzation, is carried out on a left-right image pair comprising workpieces; Canny edge extraction is carried out on a binary image, and discrete edge points are obtained through uniform sampling; according to the histogram distribution of the shape context, a candidate matching point collection is determined, a similarity measurement and calculation formula is improved, and rough matching of the shape context is carried out; the gradient direction histogram features are used for fine matching of the gradient direction histograms; and left-right consistency is introduced for calibrating and removing error matching point pairs. In the condition of meeting real-time performance requirements, the original shape context matching precision and the matching robustness can improved, and a foundation is laid for realizing quick and accurate workpiece 3D positioning subsequently.

The invention discloses an automatic k adjacent local search heredity clustering method for a graphic image, which mainly overcomes the defect of the conventional automatic clustering algorithm that the local optimization is easy to cause. The automatic k adjacent local search heredity clustering method comprises the realization steps of: (1) detecting an outline of an image by utilizing a canny edge detector; (2) describing the outline of the image by utilizing a shape context method and calculating a matched cost matrix of an outline point; (3) matching the outline point by utilizing a dynamic programming method according to the matched cost matrix; (4) converting the matched outline point by utilizing a procrustes analysis method; (5) representing the converted matched outline point and measuring an edit distance between character strings; (6) calculating the distance between the images according to the edit distance of the character strings; (7) clustering the images by utilizing a heredity automatic clustering method; and (8) carrying out k adjacent local search on groups of a heredity method. The automatic k adjacent local search heredity clustering method for the graphic image has the advantages that overall optimization is easy to achieve and an accurate clustering number can be found out.

The invention discloses an eSC and HOG-based adaptive HMM sign language identifying method which includes the steps of S1: conducting density-based sampling for the track of a sign language, extracting shape context characteristic, and combining with a pyramid processing method to obtaining eSC characteristics including space and time information, S2: for hand shape characteristics, extracting a rectangle frame containing a hand from image data and extracting HOG characteristics in the frame to realize description of the hand shape characteristics, S3: establishing an adaptive HMM model based on extracted eSC characteristics and HOG characteristics, and S4: for the data to be identified, extracting eSC characteristics and HOG characteristics through S1 and S2, and identifying the eSC characteristics and HOG characteristics extracted from the data to be identified through the adaptive HMM model to obtain the identifying model. Through the method, the correct rate of sign language identification is improved substantially.

The invention provides a vision-based multifunctional method for identifying hand gestures, which comprises the following steps of: firstly, extracting a hand part region from the background by background subtraction; secondly, carrying out fingertip detection on the extracted hand part region to determine the number of single fingertips contained in the hand gestures; meanwhile, acquiring the outline of the hand part region and describing the outline of the hand part region through improved shape context descriptors; afterwards, introducing a directed acyclic support vector fleet to classify the extracted improved shape context descriptors; after classifying, comparing the number of fingers of a hand form, to which the improved shape context descriptors belong, with the detection result of the fingertip, if the number is matched with the result, outputting the final judgment result; and otherwise, refusing the judgment. The vision-based multifunctional method can be better applied to man-machine interaction, mobile equipment and sing language input.

The invention provides an automatic classification method for traditional moire patterns, mainly solves the problem of low manual classification efficiency of moire patterns, and realizes automatic classification of the moire patterns through moire pattern preprocessing, feature extraction and clustering processing. An implementation process includes the steps of: (1) preprocessing moire images, including three steps of unifying an image size, removing background noise and thinning moire image lines; (2) for shapes between moire images, main features of which are lines, adopting a shape context descriptor (SC) algorithm to extract features of the moire images, and obtaining initial similarity of the moire images through shape context distance; (3) optimizing a similarity matrix through an improved neighbor relation transfer algorithm; and (4) using the optimized similarity matrix as an input matrix of an MEAP algorithm to perform MEAP clustering processing, thereby realizing automatic classification. A clustering result shows that compared with SIFT-MEAP and ED-MEAP algorithms, the automatic classification method provided by the invention is higher in clustering accuracy, and a clustering effect is more ideal. At the same time, the automatic classification algorithm for the moire patterns proposed by the invention has very good reference significance for clustering analysis of other traditional art patterns.

The invention discloses an image characteristic matching method based on improved shape contexts, and belongs to the medical computer technology field. The image characteristic matching method comprises steps that required SIFT characteristic points are extracted from two to-be-matched images, and according to an SIFT algorithm, rough matching is carried out; the improved shape contexts of every matching point with respect to other points are calculated; whether similarity between the relative relation among the matching points of the first image and the relative relation among the matching points of the second image exists is observed, and then whether the characteristic matching of the two images is correct is determined. The improved method is advantageous in that under a precondition of not changing dimensions of characteristic descriptors, the improved shape contexts are used to eliminate wrong matching phenomena, and therefore matching precision is improved to a great extent, and at the same time, the method is resistant to noise influences and image geometric changes.

The invention discloses a printed circuit board image registration method based on the shape context of the mass center of a communicated region, and belongs to an image registration method based on characteristics. An interest point serves as the coordinate of the mass center of the communicated region of an image. The coordinate of the mass center is characterized in two aspects, firstly, with the coordinate of the mass center as a center, the shape context of the communicated region where the coordinate of the mass center is located describes the shape characteristics of the communicated region, and therefore the shape context is called as a shape descriptor; secondly, the coordinate of the mass center is opposite to the shape context of the coordinates of the mass centers of other communicated regions in the image where the communicated region is located, the shape context describes the relative positions of the coordinate of the mass center and the communicated region where the coordinate of the mass center is located in the whole image, and therefore the shape context is called as a relative position descriptor. Through the shape descriptor and the relative position descriptor of the coordinate of the mass center of the communicated region, accurate matching between a communicated region of reference images and a communicated region of images to be registered is achieved, transformation parameters between the reference images and the images to be registered are estimated according to the coordinates of the mass centers of the communicated regions which are accurately matched, and finally the images to be registered are accurately registered.

The invention discloses a positioning method based on a shape context feature. The method comprises the steps of (S1) performing image acquisition on a workpiece and performing preprocessing, (S2) performing threshold segmentation on a workpiece image to extract a contour image, (S3) identifying a workpiece contour from a detected contour, (S4) carrying out stereo matching on left and right diagrams, and (S6) obtaining three-dimensional coordinate information of a grab point. According to the method, a shape context feature detection algorithm is used to recognize the workpiece contour in thedetected contour, the combination of a histogram and Harris corner points is used to generate and match the feature, the matching speed is accelerated, the precision is high, the method has certain adaption ability to environmental change, and the workpiece can be accurately positioned.

A blood vessel registering method based on multi-sequence medical images includes the following steps that 1, the multi-sequence medical blood vessel images are obtained, one sequence image is selected as a registering reference image, and the other sequence images serve as floating images; 2, the multi-sequence medical blood vessel images are denoised; 3, shape context description of blood vessel edges is carried out; 4, edge points are matched; 5, mismatching is filtered out; 6, edge correction and interpolation are carried out, and the blood vessel wall is evolved iteratively according to the matched points, so that the blood vessel outline energy function E (pi,qj) of the reference image and the floating images after shape changing reaches the minimum, according to an edge correction result, a spline interpolation method is adopted to carry out interpolation on partial blood vessel regions of the floating images, and finally a registration result is obtained. The blood vessel registering method based on the multi-sequence medical images is high in accuracy.

The invention discloses a human body shape identification method based on shape context suitable for garment customization. The method includes: step A1 the shoulder, back, chest and buttocks of the human body are induced into different body shapes and their characteristic parameters; A2) calculatING that shape context descriptor of the feature contour of each part to establish a body shape feature database; Step B1, acquiring a face image of a customer; B2, read that contour figure data of the human body extracted from the image, and calculate a shape context descriptor of the actual characteristic contour of the designated part; B3, compare that obtained somatotype result; C1, obtaining that actual size; C2, formulating a modification rule; C3, adjusting and modifying that specific partof the standard garment template; C4: making pattern design and outputting. The invention has the following advantages: the method is scientific and feasible, the image can be obtained remotely, the figure identification is intelligently completed, the actual size is measured, and the pattern output of the customized garment is completed.

The invention discloses a three-dimensional coordinate positioning method based on machine vision. The method comprises the steps of (S1) performing image acquisition, preprocessing and contour extraction on a workpiece, (S2) performing distance normalization processing on a feature vector, (S3) determining a candidate matching point set, (S4) obtaining an initial matching point set, and (S5) removing a mismatched point pair. According to the scheme of the invention, a shape context is introduced into the stereo matching of a binocular visual workpiece image, a centroid constraint condition isintroduced, the shape context and a gradient direction histogram are fused to form a joint similarity measurement, an RANSAC is introduced to eliminate mismatched points, high matching precision anda fast matching speed are achieved, and the rapid and accurate workpiece depth information extraction and 3D positioning requirements of binocular vision in the industry are satisfied.

The invention provides a pedestrian recognition, positioning and counting method for a video. The method achieves the recognition, positioning and counting of pedestrians in a video through the matching angle of the shape context of each point in a video image, and carries out the matching with sample points in each image frame of the inputted video to be recognized through building a standard feature library containing the shape context features of sample points, so there is no need to carry out a large amount of training for the standard feature library and only the standard feature library containing typical body postures needs to be built. Moreover, the method enables a body image to be segmented into the sample points even if a human body is complex. Because each sample point is matched with the corresponding sample point in the standard feature library, the method is higher in recognition precision, reduces the recognition error, is wider in application range, and can be further applied to the main body recognition and tracking in the video.

The invention provides an automatic door which comprises an automatic door body and a video monitoring device mounted on the automatic door body. The video monitoring device comprises an image collecting module, an image initial processing module, a filtering module, a background modeling module, a background reducing module and a feature matching module. The image collecting module is used for collecting video monitor images. The image initial processing module is used for carrying out movement target initial detection on initial video image sequence information and outputting effective video image sequence information containing a movement target. The filtering module is used for receiving the effective video image sequence information and filtering the background shape of the effective video image sequence information. The background modeling module is used for establishing a filtered background model, the background model is composed of shape context histograms of N expression peripheral points with weights. The background reducing module is used for carrying out attribute classification on the peripheral points on the current frame of image and reducing the peripheral points belonging to the background. According to the automatic door, false detection on the movement target caused by background noise and vidicon shaking can be reduced to the maximum degree, and good real-time performance is achieved.