70 results about "Shape vector" patented technology

A hand washing monitoring system (1) comprising a camera (2), a processor (4), the processor being adapted to receive from the camera images of hand washing activity. The processor analyses mutual motion of hands to determine if the hands mutually move in desired poses, and if so, the durations of the patterns; and generates a hand washing quality indication according to the analysis. The processor extracts information features from the images and generates feature vectors based on the features, including bimanual hand and arm shape vectors, and executes a classifier with the vectors to determine the poses. The processor uses edge segmentation and pixel spatio-temporal measurements to form at least some of the feature vectors.

In a quantization step information encoding unit, an average value of the quantization step information is found in an approximate shape extraction unit (20), first of all, from one set of a given number of unitary quantization units to another. In an approximate shape encoding unit (21), the approximate shape information is vector-quantized. In a residual signal computing unit (22), the residual signals between the quantization step information and the quantized approximate shape vector are computed. In a residual signal encoding unit (23), the residual signals are variable length encoded, and the so encoded residual signals and the vector quantized approximate shape information are output.

An object of the present invention is to provide a map matching method capable of performing a process operation in a high speed. In the map matching method of the present invention, while employing road network data in which road networks are formed in a plurality of hierarchical layers “a”, “b”, “c”, and a road network of an upper-grade hierarchical layer is established by such a road network which is extracted from a road network having a lower-grade hierarchical layer lower than the road network having the upper-grade hierarchical layer, a road network of the uppermost-grade hierarchical layer “a” is matched with a shape vector indicative of a shape of a objective road in a beginning stage; and when such a candidate road which is matched with the shape vector cannot be obtained in the road network of the uppermost-grade hierarchical layer “a”, since such a road network having another hierarchical layer “b”, or “c” is matched with the shape vector, the objective road is identified. In the most cases, as to a objective road of traffic information and the like, the objective road can be identified by executing a map matching operation with employment of the road network having the uppermost-grade hierarchical layer “a”, and the map matching operation can be carried out in a high speed.

A statistical facial feature extraction method is disclosed. In a training phase, N training face images are respectively labeled n feature points located in n different blocks to form N feature vectors. Next, a principal component analysis (PCA) technique is used to obtain a statistical face shape model after aligning each shape vector with a reference shape vector. In an executing phase, initial positions for desired facial features are firstly guessed according to the coordinates of the mean shape for aligned training face images obtained in the training phase, and k candidates are respectively labeled in n search ranges corresponding to above-mentioned initial positions to obtain kⁿ different combinations of test shape vectors. Finally, coordinates of the test shape vector having the best similarity with the mean shape for aligned training face image and the statistical face shape model are assigned as facial features of the test face image.

The present invention provides a method and apparatus for compressing data of position information such as a road on a digital map in order to realize an accurate transmission of the position information in a small volume of data. A shape vector of an object road segment is resampled, and as this occurs, a sampling point is represented by quantized angle information. By selecting preferably the sampling point and an angle resolution of the quantized angle information, an error in position information can be reduced. For example, in a method for compressing position information by setting appropriately an angle resolution δ which constitutes a quantization unit of an angle, the angle resolution is set according the length of a resampling length L which regulates an interval of resampling, a road shape contained in a digital map is divided into one or a plurality of segments, road shapes in the segments are resampled by a constant resampling length L, and the position of the road shapes is represented by a data string of quantized angle information indicating the position of a sampling point, and data of the data string are variable length coded for transmission. By making the angle resolution δ and the resampling length L interlock with each other, the angle resolution is made small when the resampling length is long, whereas when the resampling length is short, the angle resolution is increased, whereby position information such as road can be transmitted accurately in a small volume of data.

The invention relates to an adjustable ring-shaped vector beam generating system. The prior art has high incident beam requirements, and cannot realize the ring-shaped vector beam and the beam size adjustment. In the invention, the incident beam sequentially passes through a beam expanding shaper, an external reflection circular conical surface reflector, a cylindrical curved polarizer, an internal reflection circular conical surface reflector, and two halfwave plates positioned in parallel; after a light source emergent beam passes through the beam expanding shaper and the external reflection circular conical surface reflector, a radial propagating beam is formed; a ring-shaped rotation vector polarized beam is formed by the reflection of the internal reflection circular conical surface reflector after the radial propagating beam passes through the cylindrical curved polarizer; the polarization state is changed by adjusting the optical axis angle of the two halfwave plates through a rotary motion part; and the transverse size of the beam is changed by adjusting the space between the external reflection circular conical surface reflector and the internal reflection circular conical surface reflector through an axial motion part. The invention has the advantages of low incident beam requirements, ring-shaped vector beam output, convenient beam polarization state and transverse size adjustment, etc.

The objective of the present invention is to provide a relative location data correction apparatus, a relative location data correction method and a relative location data correction program that can cope with the shifting of a relative location between different digital map databases, and can accurately display a desired point.

The relative location data correction apparatus of the invention includes: a transmission apparatus (100a), for transmitting to a reception apparatus (200a) event information, such as that for a traffic accident or traffic congestion, that is prepared based on a digital map database (101); and the reception apparatus (200a), for employing the event information received from the transmission apparatus (100a) to display an event occurrence point on a map provided by a digital map database (207), and for correcting the location of the event occurrence point before being displayed. The transmission apparatus (100) calculates the total length of a road section, and transmits, to the reception apparatus (200a), shape data expression information that includes the total length. The reception apparatus (200a) calculates the total length of the road section, and employs the two total lengths to correct the relative location of the event occurrence point.

Systems and methods for shape registration are described. In one aspect, training shape vectors are generated from images in an image database. The training shape vectors identify landmark points associated with one or more object types. A distribution of shape in the training shape vectors is represented as a prior of tangent shape in tangent shape space. The prior of tangent shape is then incorporated into a unified Bayesian framework for shape registration.

The invention discloses an improved CT image aorta segmentation method based on an active shape model. A segmentation and extraction method combining a Support vector machine (SVM) and an active shapemodel (ASM) are used to segment and extract the aortic region from the CT images of patients with aortic dissection accurately, which solves the problem of error caused by the large difference between the model and the actual segmentation target in the existing algorithm. The process is as follows: (1) selecting samples from aortic patient CT images to construct a training set and marking the characteristic points of an aortic region in the training set; (2) constructing a shape vector from the feature points of the sample marks and performing normalized registration; (3) reducing the dimension of the vector to determine the main sample components and constructing a statistical shape model; (4) performing gray-scale sampling ofa square matrix centered on feature points to establish a texture model; (5) constructing a support vector machine classifier in the training process of the model; (6) calculating the probability of the marked point set to the contour of the target, and findingthe best matching position.

The invention discloses a method for detecting and positioning key characteristic points in face recognition. The method comprises the steps of preprocessing a face image; presetting 22 key characteristic points; according to the calibration sequence of the preset key characteristic points, artificially calibrating the key characteristic points in the training samples for forming a set of shape vector training samples; constructing a global shape model and a local texture model; performing searching positioning on the key characteristic points; optimizing positioning of the key characteristic points; automatically calibrating the key characteristic points and performing normalized outputting of the key characteristic points. Because the method reduces a sampling range, sampling efficiency in subsequent extracting the key characteristic points can be effectively improved.

An age-converting method of human face based on average face and ageing scale map includes dividing human face to be shape vector and vein vector based on dense feature presentation, applying human face cartoon synthetic technique to finalize young conversion of specific human face by calculating difference between specific human face to average human face, combining human face image vector modeling with human face image ageing conversion for realizing human face image ageing analog in accordance with physiological feature.

The invention discloses a human face image age changing method based on average face and aging scale map, comprising the steps of: dividing the human face to be shape vector and vein vector with dense feature presentation; using a human face cartoon synthetic technique to complete young changing of specific human face by calculating difference between the specific human face and the average human face and further reducing the difference; and combining human face image vector modeling with human face image ageing changing for realizing human face image ageing analog in accordance with physiological feature. According to the changing of human face shape feature caused by aging, some feature line of the human face shape is modified so as to realize shape ageing analog; and for the change of vein feature, an aging scale map is prepared in a way of scale map, according to vein mapping, the vein feature aging analog of the human face shape is realized. The method fully considers the unified change of the shape and the vein when human face age changes, thereby greatly increasing the third dimension of synthetic image.

A method for generating Pairwise Active Appearance Models (PAAMs) that characterize shape, appearance and motion of an object and using the PAAM to track the motion of an object is disclosed. A plurality of video streams is received. Each video stream includes a series of image frames that depict an object in motion. Each video stream includes an index of identified motion phases that are associated with a motion cycle of the object. For each video stream, a shape of the object is represented by a shape vector. An appearance of an object is represented by an appearance vector. The shape and appearance vectors associated at two consecutive motion phases are concatenated. Paired data for the concatenated shape and appearance vectors is computed. Paired data is computed for each two consecutive motion phases in the motion cycle. A shape subspace is constructed based on the computed paired data. An appearance subspace is constructed based on the computed paired data. A joint subspace is constructed using a combination of the shape subspace and appearance subspace. A PAAM is generated using the joint subspace and the PAAM is stored in a database.

A method for quantizing prediction memory during fast recovery coding is disclosed. A best shape vector that describes prediction memory for a current frame is quantized. It is determined whether to send the quantized best shape vector. The quantized best shape vector is sent based on the determination. An encoded current frame is sent.

The invention discloses a numerical control machine tool sensitive-link identification method based on a modal mass distribution matrix. The modal mass distribution matrix of a numerical control machine tool is constructed with the modal mass of the numerical control machine tool and a modal shape vector, and as a result, the sensitive link of the machine tool structure can be obtained. The method is characterized by comprising the steps of (1) carrying out a modal experiment of the numerical control machine tool to obtain the modal mass of the numerical control machine tool and the modal shape vector, (2) constructing the modal mass distribution matrix of the numerical control machine tool according to the obtained modal mass and the modal shape vector, and (3) obtaining a modal mass distribution curve of all components of the machine tool by virtue of the modal mass distribution matrix, and further identifying the sensitive link of the machine tool. The method is capable of determining the sensitive link of the machine tool structure so that the weak condition of the machine tool structure is quantized, and therefore, the previously mentioned problems encountered in the process of evaluating the sensitive link of the machine tool structure by an existing method are solved.

The invention discloses a face recognition method, system and a storage medium. According to the position and size of human face, obtaining the human face image, and preprocessing the human face image; extracting features of the preprocessed face image; Training the prior data to generate a feature template; taking The feature template to match the extracted face image features, and obtaining theface recognition results. The invention adopts a face recognition method based on feature template matching, a stereo-adjustable shape vector training sample is established based on the characteristics of human face, After locating the human face position, the human face image is matched through the feature template, which solves the problem that the prior art is easily affected by the observation angle, occlusion and expression change in the process of face recognition, improves the accuracy of face recognition, and can be widely used in the field of depth learning technology.

The invention discloses a facial expression feature extraction method based on a point distribution model. The method includes the specific steps of setting up the human body facial shape model through coordinates for marking human body facial image facial feature points, selecting relatively-ideal shape vectors as initial samples, enabling other vectors to correspond to the initial samples in shape till the difference between two adjacent average shape vectors is smaller than a certain value, and conducting calculation through a Gabor wavelet kernel function and conducting sampling to achieve fine mapping of the facial feature points of a set of feature points with different frequencies and phases. By means of the method, according to the facial expression feature extraction method based on the point distribution model, the elastic graph matching of human face recognition is conducted through a Gabor wavelet conversion coefficient, ideal facial expression features which are not affected by the samples are obtained, and the method has the obvious advantages for extracting features which are kept unchanged under uncorrelated deformation, insensitive to noise and high in category distinction degree.

The invention relates to a safety assessment method for a heavy load structure of a heavy type construction machine. The safety assessment method includes the concrete steps that modal shape vectors and stiffness matrixes are used for determining corresponding structural element modal strain energy and corresponding overall structure modal strain energy, wherein the modal shape vectors and the stiffness matrixes are generated before and after structural damage; element modal strain energy ratios before and after structural damage is determined by adopting the ratio of the element modal strain energy to the overall structure modal strain energy; the strain energy damage tolerance ratio degree DESR ij is determined through the ratio of element modal strain energy before structural damage to element modal strain energy after structural damage. When the modal strain energy damage tolerance ratio degree DESR ij approaches zero, it shows that the j element is not damaged, and the larger the modal strain energy damage tolerance ratio degree DESR ij is, the more serious the damage of the j element is damaged. According to the method, the modal strain energy damage tolerance ratio degrees of all the elements are sequentially calculated so that the positions and the degrees of structural damage can be obtained. Existence of structural damage of the heavy load structure of the heavy type construction machine and the positions of structural damage can be fast and accurately recognized, and the problem that structural damage of the heavy load structure of the heavy type construction machine is not prone to being recognized is solved.

The invention provides a three-dimensional torus-shaped vector hydrophone, which comprises a neutral buoyancy torus, an elastic element, a hanger loop and a rigid strut, wherein the neutral buoyancy torus and the rigid strut are flexibly connected in a central fixation manner through the elastic element and the hanger loop. The invention breaks through limitation that the original resonant type vector hydrophone is spherical or columnar generally, underwater acoustic vector signals can be accurately obtained, and the three-dimensional torus-shaped vector hydrophone has the advantages of small dimension, high sensitivity and flexibility in use.

The entire block of the transmission area of information such as shape vector data representing road shape and event information is divided into partitions of a plurality of layers and block definition is made so that the data amount in each partition is approximately equal. The transmitting side generates and transmits shape vector data, event information, shape vector index data including reference data indicating correspondence between the block definition and each partition, and a shape vector list indicating shape vector data contained in each partition. The receiving side calculates partitions necessary for processing and extracts shape vector data contained in the partitions based on the shape vector index data and the shape vector list, then performs shape matching of the extracted shape vector data and identifies the section and the position in the section on the map to which the event information corresponds.

The invention discloses an aortic contour segmentation algorithm based on an active shape model, which mainly overcomes the blurring or discontinuity of the aortic edge caused by the traditional method through the shape constraint, avoids the interference of other tissues, and realizes the accurate segmentation of the aortic arch part and the descending aorta part contour. The realization processis as follows: (1) marking feature points of a training set sample and performing normalizing and registration treatment on the shape vectors to establish a point distribution model; (2) sampling thegray level along the normal direction of the contour boundary of the training sample set to construct the gray level texture model of the training set; (3) searching for the best matching point of each marker point iteratively, and then uniformly performing constraining by the active shape model until the aortic contour converges.