128 results about "Movement recognition" patented technology

Virtually fraud-proof authentication for wireless mobile financial transactions. A mobile telephone, genius card or dongle may capture multiple biometrics, such as fingerprint, voiceprint, retinal recognition data, facial biometric data, ear biometric data and bio print, iris scanning data, pupil movement recognition data, and send segments of each of them to a server computer that re-constitutes them and stores them for matching during an attempted financial transaction. A purchasing individual has multiple biometrics read by separate readers integrated into a single hand-held device. An automatic process of the server computer has previously sent a command packet to the hand-held device defining percentage and portion of each biometric to select and the sequence and channel of transmission for matching along with other out of band security data. During the charging phase, the hand-held device may utilize a slidable antenna for swiping a magnetic strip of a charge card reader or may display credit card data for a limited amount of time.

A system for detecting, tracking, displaying and identifying repetitive movement, including a sensor configured to sense movement, and in particular static acceleration, along at least a first horizontal axis, and ideally about a second horizontal axis, with respect to a vertical axis and a processor to generate output signals therefrom for audible and visual display of information that can include movement identification, movement patterns, and to further include elapsed time, start and stop times, breathing patterns, and variations thereof from a reference.

The invention provides a human body movement recognition method based on convolutional neural network feature coding and mainly aims to solve the problems of complicated calculation and low accuracy in the prior art. According to the implementation scheme, TV-L1 is utilized to obtain a video light steam graph; convolutional neural network coding, local feature accumulation coding, dimension-reducing whitening processing and VLAD vector processing are sequentially performed in a video space direction and a light stream movement direction, and space direction VLAD vectors and movement direction VLAD vectors are acquired; and information in the video space direction and information in the light steam movement direction are merged to obtain human body movement classification data, and then classification processing is performed. According to the method, convolutional features are subjected to local feature accumulation coding, so that the recognition rate is increased when complicated background data is processed, and the calculated amount is reduced; the features acquired by fusing video VLAD vectors and light stream VLAD vectors has higher robustness to environmental changes, and the method can be used for performing detection and recognition on human body movement in a monitoring video in areas such as a community, a shopping mall and a privacy occasion.

The invention relates to a method for implementing realistic game, which comprises the following steps of: (1) establishing a human body skeleton model; (2) establishing a game movement library under an offline state, establishing movement libraries respectively according to game items, and performing multi-frame movement decomposition on a single semantic movement;(3) calibrating a binocular camera to acquire parameters of the binocular camera and polar calibration; (5) background modeling; (6) selecting an interactive characteristic label; (7) foreground partitioning; (8) initializing the information of the characteristic label; (9) detecting human face and skin color; (10) multi-target tracking; (11) completing sparse stereo matching; (12) acquiring a 3D skeleton; and (13) matching with the movements in the offline movement libraries to realize movement recognition, combining single-frame image analytic matching with multi-frame image analytic matching to obtain the semantic movement, transferring the semantic movement to a game executing unit to implement the function of realistic game. Being stronger in interactivity and reality as well as simpler and more convenient in operation, the method of the invention is a game implementation method with low cost and is more suitable for being extensively accepted by general people.

The invention discloses a device of assisting a user in body building, which comprises a memory, a display unit, a keyboard, a loudspeaker, a power module, and also a shooting module which is used for converting external scenes into corresponding digital image files; an image recognition module which is used for recognizing human body movements in the images and extracting features of human body movements; and a main processing module which is used for controlling movements of every module and outputting comparing results. The device of the invention is equipped with a human body movement recognition system, compares human body building movements with standard movement templates, can provide guides like a body building coach for users at any time and any place, is used for pointing out non-standard places in user body building movements and helps users achieve ideal body building effects.

The invention discloses an intelligent indoor illumination control method based on dynamic video identification, which comprises the following steps of: (1) arranging an information acquisition module, and collecting an information source; (2) arranging a function control device; (3) arranging a light control device; (4) connecting all modules in the above steps mutually, electrifying and enabling the modules to work; (5) processing a signal by a central processing module, and outputting a light control signal; and (6) according to the received control signal, controlling a controlled luminaire by the light control device. The invention also discloses a system which implements the method. By a camera of the system disclosed by the invention, the intelligent control of a lighting system is effectively implemented by means of character recognition, movement recognition, gesture control, illuminance induction and the like so as to comfortably, flexibly and efficiently control illumination in a humanizing mode, so that the method is applicable to various indoor illumination application environments.

A cleaning robot includes a main body, a moving assembly to move the main body, a cleaning tool provided at a bottom part of the main body to collect foreign substances on a floor, an imager to collect images around the main body and a controller to recognize motion of a hand by performing image processing of the collected images, identify a control command corresponding to the motion of the hand, plan a moving direction and a moving distance of the main body as movement information based on the control command, and control operations of the moving assembly and the cleaning tool based on the planned movement information. Since the user directly controls movement of the cleaning robot, it is possible to improve interactivity between human and cleaning robot, reduce the user's labor and increase convenience.

A movement recognition method and a user interface are provided. A skin color is detected from a reference face area of an image. A movement-accumulated area, in which movements are accumulated, is detected from sequentially accumulated image frames. Movement information corresponding to the skin color is detected from the detected movement-accumulated area. A user interface screen is created and displayed using the detected movement information.

Virtually fraud-proof authentication for wireless mobile financial transactions. A mobile telephone, genius card or dongle may capture multiple biometrics, such as fingerprint, voiceprint, retinal recognition data, facial biometric data, ear biometric data and bio print, iris scanning data, pupil movement recognition data, and send segments of each of them to a server computer that re-constitutes them and stores them for matching during an attempted financial transaction. A purchasing individual has multiple biometrics read by separate readers integrated into a single hand-held device. An automatic process of the server computer has previously sent a command packet to the hand-held device defining percentage and portion of each biometric to select and the sequence and channel of transmission for matching along with other out of band security data. During the charging phase, the hand-held device may utilize a slidable antenna for swiping a magnetic strip of a charge card reader or may display credit card data for a limited amount of time.

A system of data entry on a surface coated with touch sensitive material is provided. This surface is referred herein as touch panel (36). This system is comprised of methods that use touch movement recognition in one of two directions up (44) and down (43) or left (54) and right (53). The touch movements are mapped to a sequential list of basic elements of a language (31). Basic elements can be alphabets (32), numerals (33), ideograms (34) and words (35). Cancellation and selection commands are optionally mapped to left (54) and right movements (53). Selected basic elements are passed onto applications in context of data entry.

The invention is suitable for the technical field of pattern recognition, and provides a human movement recognition method and device. The method comprises the steps: obtaining a depth image sequence, carrying out the conversion of the depth image sequence, and obtaining a corresponding depth movement sequence; carrying out the dividing of the depth movement sequence in the time dimension and space dimension, and obtaining a plurality of movement historical cubes and a plurality of corresponding space cube subblocks; calculating corresponding characteristic vectors corresponding to the movement historical cubes according to the space cube subblocks, and obtaining the characteristic vector of the depth movement sequence through combining the characteristic vectors corresponding to the plurality of movement historical cubes; and carrying out the model training and testing through employing an SVM (support virtual machine) according to the characteristic vector of the depth movement sequence, so as to obtain a recognition result of human movement. The method solves a problem of self-sheltering in a conventional human movement recognition method, improves the description capability for human movement, and improves the recognition accuracy of human movement.

The invention relates to the field of human body movement recognition, and provides a swimming attitude measuring method based on a wearing-type sensor. The method comprises the steps that a moving signal and a magnetometer signal of limbs when a body stands and swims are collected by using the wearing type sensor respectively; the initial attitude of the limbs is obtained, correction is conductedon the initial attitude, and the corrected initial attitude is obtained; quaternion of a sensor coordinate system and a geographical coordinate system can be obtained through a gravity vector direction and a magnetic field north magnetic pole direction, and the quaternion is utilized to perform calibration on the body initial attitude and a biomechanic model, a complementary filtering algorithm is utilized to eliminate sensor error, and body swimming attitude updating is conducted by using the moving signal and the magnetometer signal in the swimming process. By means of the method, swimmingbody attitude can be captured, and the attitudes of swimmers are comparable through kinematic analysis.

The invention discloses a human body movement recognition method based on surveillance isometric mapping, and belongs to the field of pattern recognition and computer vision. The human body movement recognition method comprises the following steps; S1, performing foreground extraction through codebook method for the video to acquire a binarized human body foreground image; S2, performing morphology processing and normalization for the human body foreground image to acquire a human body silhouette image; S3, performing periodization analysis for the human body silhouette image sequence, wherein each video is represented by a series of key frames comprising a complete movement period; S4, performing vectorization for the key frames of the human body silhouette image, and performing characteristic dimension reduction through surveillance isometric mapping; S5, recognizing the characteristic after dimension reduction through the nearest categorizer according to Hausdorff distance. The human body movement recognition method breaks through the limitation of the conventional algorithm, and reduces complexity of the algorithm while increasing the categorizing accuracy, thereby being more feasible in practical engineering application.

The invention provides a vehicle anti-theft system based on human face expression and movement recognition, and belongs to the technical field of image processing and mode recognition. The technical problems of vehicle anti-theft effect and the like are solved. The vehicle anti-theft system based on human face expression and movement recognition comprises a memory module, a face recognition module, an expression and movement recognition module, a communication module, a controlling module and the like. The vehicle anti-theft system based on human face expression and movement recognition has the functions of preventing false alarming, preventing deception and preventing holding.

The invention discloses a footstep movement identification method based on a triaxial acceleration sensor. The method comprises the following steps: S1, reading detection data output from the triaxial acceleration sensor by a microprocessor, and processing corresponding data selected according to a body kinematic state; S2, drawing an acceleration change curve according to the selected data by the microprocessor, detecting the peak of wave in the acceleration change curve and determining the falling time of the foot according to the peak of the wave; S3, calculating the falling posture of the foot by the microprocessor; S4, judging the current body kinematic state by the microprocessor and setting. According to the method of the invention, the falling posture of the foot of people during movement can be identified, so that the convenience is brought to a user to know the self-postures of walking and running.

A multiple exercise activity recording system in which each piece of equipment, exercise movement exercise area, fitness activity or biometric in a facility where exercises are pre-formed is assigned to an exercise identification module. Each exercise identification module is linked or coupled to a terminal with a display and a manual input device. Each user is assigned a personal identification device that is presented or inputted into the terminal before or after the exercise equipment or exercise area is used. After exercising, the user activates the exercise identification module assigned to the exercise equipment or to the exercise area. Upon activation, one or more sub-routines automatically run in the terminal. Prompts presented on the display then request user input of information pertinent to the exercise activity. The inputted information may be stored in the terminal or immediately transmitted to a permanent member data file on a local or remote server.

The invention discloses a human body behavior action fast recognition method and device. The recognition method includes the steps of extracting contours, acquiring key frames, calculating characteristic operators, setting thresholds and the like, and human body behavior actions are classified according to comparison analysis between the characteristic operators and the thresholds; the recognition device is composed of a movement contour acquiring unit, a behavior model establishing unit and a movement recognition unit. The method and device has the advantages of being simple, efficient, reasonable in use and the like. The matching calculation amount can be reduced, the matching calculation time can be shortened and the real-time performance and the accuracy of behavior action recognition can be improved.

The invention discloses a human movement recognition method based on multistage characteristics. The human movement recognition method based on the multistage characteristics includes the steps of extracting a space-time interesting point of each frame of image in an original video, classifying the obtained space-time interesting points according to the positions of the human skeletons, obtaining human areas, establishing multiple stages according to the obtained human areas, clustering the space-time interesting points included in the human areas in the stages in a training set according to a clustering algorithm, obtaining corresponding dictionaries, extracting word bag characteristics of the human areas in the training set and a test set through a word bag model, fusing the word bag characteristics of the obtained three stages, and learning and judging the human movement through a hidden conditional random field model for movement recognition. By means of the method, local characteristics can be represented more effectively, and tests verify that the method improves the accuracy of movement recognition.

The invention discloses an achieving method of a human movement recognition training system. The method includes the following steps of (1) collecting training information, (2) normalizing sample data, (3) extracting a characteristic three-dimensional vector set, and (4) recognizing and identifying motions. According to the method, a reduced precision discretization coordinate algorithm is used for simplifying operation information representation and reducing calculation amount of a characteristic extraction algorithm, and meanwhile a multi-level feature matching algorithm is used for accelerating the recognition speed. The achieving method is simple and reasonable in logical design and algorithm, feasible, reliable and easy to achieve.

The invention discloses a circuit breaker mechanical characteristic on-site live-line test method based on a normalized cross correlation image pyramid sector matching (NCC-P-S) optimization algorithm. The method comprises the following steps: installing a circuit breaker mechanical characteristic parameter live-line test specific device, adjusting device parameters and starting the test process;capturing a high-speed image sequence associated with movement of a moving contact by utilizing a high-speed camera, and determining some characteristic part on a main shaft or a crank arm of a circuit breaker body as a motion recognition target; identifying the moving object through the NCC-P-S optimization algorithm to obtain movement trajectory of an insulated connecting rod or the main shaft;modifying small vibration of the center of circle through a multi-target linkage center location method, and calculating main shaft rotation angle through a multi-target linkage weighted discrimination method; and finally, calculating mechanical characteristic parameters of a circuit breaker through association relation between the main shaft and the moving contact. The method is a new non-contactcircuit breaker mechanical characteristic test method, can solve the problem of sensor installation, and can realize live test of circuit breaker mechanical characteristic parameters.

The invention relates to a front man-machine interaction voice recognition method and a front man-machine interaction voice recognition system based on computer vision assistance. According to the method and the system, video signal input is added in the traditional voice recognition flow, and then the simultaneous recognition with the voice signal is realized; voice assistance is carried out in human face recognition and human face lip movement recognition, and whether a target to be recognized is speaking or not is judged; and meanwhile, through human face recognition and auxiliary positioning, the orientation of a speaker is judged, and in addition, enhancement processing is carried out on a sound source signal in the specified direction by utilizing the corresponding orientation. Withthe method and the system, the recognition accuracy for voice command and voice input information of clients can be effectively enhanced under the specific environments, including the man-machine interaction use situations in which the clients need to face the device from the just front side, such as self-service retail terminals, bank self-service terminals and insurance self-service terminals.

The invention is applied to the technical field of televisions and provides a remote control television. The remote control television comprises a range image induction unit for obtaining image range data by inducing range images of the limbs of a user; a limb movement recognition unit for recognizing movements of the limbs of the user according to the image range data obtained in the range image induction unit; a command conversion unit for converting the movements of the limbs of the user recognized by the limb movement recognition unit into corresponding television control commands; and a command execution unit for executing the television control commands converted by the command conversion unit. Based on infrared image induction, the remote control television separates the limbs of the user from a complex environment background by obtaining relevant information of the range images and carries out recognition and command matching to the movements of the limbs of the user, and accordingly the aim that the television is controlled without using a traditional remote controller is achieved.

The invention discloses a mobile terminal movement recognition method based on infrared. According to the method, the movement direction is recognized through an infrared emitter and an infrared receiver. Compared with a traditional human body movement recognition method based on vision, installing is convenient, cost and power consumption are low, the responding speed is high, the algorithm is simple, the sensitivity is high, stability and reliability are achieved, the mobile terminal movement recognition method is very suitable for being applied to a mobile terminal, system software and application software of the mobile terminal can very conveniently achieve gesture flipping over, incoming call gesture answering, gesture unlocking, gesture dialing and other distinctive functions through an interface provided based on the method.

The invention discloses a neural network-based human body movement recognition method. The method includes the following steps of based on a video database, training N mutually independent 3D convolutional neural networks as a video feature extractor; according to the video feature extractor, training a multi-instance learning classifier; inputting a to-be-recognized video, extracting video features through the well trained network, and classifying movements by the classifier. According to the technical scheme of the invention, the influence of a large amount of noise features on a classification result is avoided. Meanwhile, the negative effect of a fixed sample length on a movement recognition result is also avoided.

The invention provides a limb movement recognition method based on the autonomous motor imagery electroencephalogram. Survey data show that when a person does the same movement, the brain can generatesimilar electroencephalogram signals, and therefore the feature signals can be extracted to achieve movement of the mechanical arm, and then the disabled can be assisted in moving. According to the method, the brain network is established to classify the actions, so that the consideration of the correlation among the regions of the brain is enhanced, and the working mechanism behind the EEG signals and the related rhythm characteristics thereof is displayed. A brand-new brain factor effect network analysis method based on a nonlinear partial directional coherence method is used for carrying out upper limb movement classification by utilizing a random forest, the classification accuracy is high, and corresponding actions are judged according to electroencephalogram signals, so that the mechanical arm moves, and the purpose of assisting the disabled in movement is achieved.

The invention discloses a behavior recognition method based on robust relative attributes. The method includes the following steps of abstracting feature vectors of all movement video samples in a video sample bank, setting a plurality of human body movement attributes corresponding to various human body behaviors, setting a relationship of every two movement video representing human body behaviors, namely, the relationship of each movement video pair, under all the human body movement attributes, treating the relationship of each movement video pair as input, conduct training through a sequencing support vector machine to obtain a training model, solving the sequencing support vector machine through a gradient descent method to obtain parameter vectors of the sequencing support vector machine and further obtain the optimal training model, and conducting human body behavior recognition on each piece of movement video to be tested through the obtained optimal training model to obtain a human body movement recognition result. Experimental results show that the behavior recognition method can improve robustness of the human body behavior recognition.

The invention provides a human movement recognition method based on cross-domain dictionary learning. The method comprises the steps that intra-class differences between training samples of a target domain are expanded through a cross-domain movement recognition framework based on a source domain data method so as to enhance the performance of an existing recognition system, and therefore annotated data obtained from a different domain are used as a source domain to be provided on the basis of manually-annotated movement information in a target domain; a domain self-adapting dictionary pair is learned through a discriminating cross-domain dictionary learning method, and data with different distributions are migrated to the same feature space so as to enable feature distribution of different domain data to be matched; corresponding annotated information between different domains is not executed so as to adapt to different kinds of migration learning in reality. According to the method, data distribution of relevant movement of a source domain database is used for adapting to data distribution of movement in the target domain, and the dictionary pair with reestablishment, discrimination and domain self-adaptability can be learned and obtained through the method.

The invention discloses a sport movement recognition method and device, electronic equipment and computer readable storage medium. The sport movement recognition method comprises the steps of tracking an executed sport process of a sport object, and acquiring original sport signals; performing conversion on the original sport signals according to a preset conversion algorithm, and extracting a plurality of features of the original sport signals from the converted signals; acquiring the occurrence probability of each prestored feature in various movement types; screening a movement type with the highest probability of simultaneous occurrence of the plurality of features from the various movement types; and acquiring a sport movement recognition result according to the screened movement type. The technical scheme provided by the invention is high in sport movement recognition accuracy, small in recognition calculation amount and high in recognition efficiency.

The embodiment of the invention provides an image processing-based implement method and an image processing-based implement device of a virtual keyboard. The method mainly comprises the steps as follows: projecting the virtual keyboard in an appointed plane via a camera and collecting an image that a user taps the virtual keyboard via the camera, and determining a target key pressed by the user according to the image and the position information of the virtual keyboard, and obtaining a releasing operation or long-pressing operation of the user to the target key by analyzing the follow-up images collected by the user to tap the virtual keyboard. According to the embodiment of the invention, by projecting the virtual keyboard to the appointed plane via the camera, and collecting the image that the user taps the virtual keyboard, and judging the pressing, long-pressing and releasing operations of the user in the virtual keyboard via a human movement recognition technology and a image analysis technique, the method and the device could achieve the convenient input of an intelligent so as to be convenient for meeting the requirement of the user.