39results about How to "Improve interaction accuracy" patented technology

The invention relates to a fingertip detection method based on Kinect depth information. A device Kinect is connected to a computer through a cable. The method is characterized by comprising the firststep of extracting a hand and acquiring a palm coordinate; the second step of performing fingertip positioning including image preprocessing and hand contour; performing joint bilateral filtering onan extracted hand area; utilizing a Douglas-Poke algorithm to approximate a specified point set, finding out a polygon fitted curve of the hand contour and drawing out a fitted curve of the hand; thethird step of utilizing a convexHull () function to search the steps, and analyzing and obtaining a convex hull point of the hand; the fourth step of obtaining the curvature at the convex hull point,and according to the difference between the curvature of the wrist and the curvature of the fingertip, setting an appropriate threshold value to remove the convex hull point at the wrist. The method can complete gesture recognition tasks in real time and accurately, improve the real-time property and accuracy of Kinect gesture recognition, and improve natural gesture interaction experience.

The invention belongs to the technical field of intelligent control, and relates to a human-computer interaction method, a control device, a controlled device and a storage medium, wherein the human-computer interaction method comprises the steps of receiving a voice signal; detecting the characteristics of a voice signal source, wherein the characteristics of the voice signal source comprise theface orientation of a user sending out the voice signal or the relative orientation of the user and the controlled device; judging whether the voice signal comprises a wake-up word or not; if the voice signal comprises the wake-up word, performing voice instruction recognition on the voice signal to acquire a voice instruction; and if the voice signal does not comprise the wake-up word, when the characteristics of the voice signal source accord with preset characteristics, entering the step of performing voice instruction recognition on the voice signal to acquire the voice instruction, wherein the preset characteristics comprise that the face of the user faces the front face of controlled equipment / the control device or the user is positioned on the front face of the controlled equipment.Therefore, the occurrence of the condition of false triggering of the controlled equipment can be effectively avoided, so that the accuracy of the man-computer interaction method is improved.

The invention discloses a voice interaction method and device. The method comprises a step of determining at least one voice data collected by multiple microphones Mic, a step of sending the at least one voice data to a service end, a step of receiving each piece of identification information and a corresponding feedback result sent by the service end, a step of carrying out semantic analysis on each piece of identification information so as to obtain a semantic recognition score of each piece of identification information, a step of determining a feedback result corresponding to any piece of identification information with a highest semantic recognition score as output data, and a step of outputting the output data. According to the embodiment of the invention, the accuracy of speech interaction is improved.

The invention discloses a fusion and interaction system for an extra-large-breadth display contact, which comprises a projector, a plurality of display screens, cameras and an interactive laser pen, wherein the projector and the display screens are connected with a projection host; the cameras and the interactive laser pen are connected with an interaction host; the display screens are spliced to form an extra-large-breadth display screen; the projector is used for projecting images onto the extra-large-breadth display screen; infrared laser emitted from the interactive laser pen is projected onto the extra-large-breadth display screen to form interactive contacts; the back of each display screen is provided with more than two cameras; the cameras simultaneously track and collect the image data comprising a laser point coordinate on a single display screen corresponding to the cameras; and the data is transmitted to the interaction host connected with the cameras; the interaction host processes the data so as to obtain an image which can reflect the laser point movement state; and the image data is fed back to the projection host and is projected on the extra-large-breadth display screen by the projector connected with the projection host.

The invention discloses a motion sensing interaction method, a motion sensing interaction apparatus, a motion sensing interaction device and a readable storage medium, wherein the motion sensing interaction method comprises the following steps: when posture change of a motion sensing interaction person is detected, acquiring posture change information of the motion sensing interaction person, andacquiring eye expression focusing information of the motion sensing interaction person; judging whether a motion sensing operation corresponding to the posture change information exists in a preset operation list; and when the motion sensing operation corresponding to the posture change information exists in the preset operation list, locating a target operation based on the eye expression focusing information and the posture change information, outputting a response result corresponding to the target operation, and displaying the response result. With the method, the apparatus, the device andthe readable storage medium provided by the invention, a technical problem of low accuracy of gesture operation response in the existing motion sensing interaction process is solved.

The invention relates to an interaction model training method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a reference interaction track, wherein the reference interaction state sequence comprises a plurality of reference interaction states which are sequenced in sequence according to an interaction sequence; selecting a current initial interaction state from the reference interaction track according to a reverse order of the interaction order; training a to-be-trained interaction model according to the current initial interaction state to obtain a current training interaction model corresponding to the current initial interaction state; and when it is determined that the current training interaction model meets the model convergence condition corresponding to the current initial interaction state, returning to the step of selecting the current initial interaction state from the reference interaction track according to the reverse order of the interaction order to obtain the updated current initial interaction state so as to continue model training until a training stop condition is met to obtain a trained target interaction model . By adopting the method, the model training effect can be improved.

One or more embodiments of the invention provide a graphic code verification method and device, and the method comprises the steps: carrying out the code scanning of a target graphic code, and generating a code scanning result for the target graphic code; judging whether the target graphic code meets a preset condition or not according to the generated code scanning result; if yes, determining whether the code scanning operation is an operation of requesting to verify whether the target graphic code belongs to an application user or not; and first interaction information representing that thegraphic code is verified successfully and the target graphic code belongs to the application user is generated. Targeted feedback information is generated on the basis of a graphic code verification result; therefore, the user can know that the feedback information is generated according to the verification requirements of the requested graphic codes for the respective applied graphic codes. Therefore, the feedback result of completing the authenticity verification of the graphic code according to the graphic code verification requirement of the user can be transmitted to the user, the interaction accuracy between the terminal and the user is improved, and the user experience is improved.

The invention discloses an information processing method and device. The method comprises the following steps: photographing a target object through an image acquisition device to obtain a target image; extracting a feature image of the target object from the target image, wherein the feature image comprises action features of the target object; analyzing the feature image by using a first model,and determining a control parameter corresponding to the feature image of the target object, the first model being trained by using multiple groups of data through machine learning, and each group ofdata in the multiple groups of data including the feature image and the control parameter corresponding to the feature image; and controlling corresponding target equipment based on the control parameters. According to the invention, the technical problems of control action delay and low efficiency caused by the fact that the existing household intelligent electric appliance controls the electricappliance according to the voice of a user or a remote controller are solved.

The invention discloses a method for the binding interaction constraint of a human body mapping posture body and a virtual rotating body. The method includes the following steps of S1, establishing avirtual reality world with a quaternion structure as the four-dimensional space; S2, establishing the human body mapping posture body of a real human body in a virtual environment; S3, representing the rotating axis of the virtual rotating body through a three-dimensional vector, and representing the rotating angle of the virtual rotating body through an angle vector; S4, setting the rotating axisas the A point and the hand as the B point, and setting the virtual rotating body to be kept on the straight line from A to B; S5, judging whether the hand keeps holding the virtual rotating body ornot, and if yes, always making the virtual rotating body move in the direction of the AB straight line to form the binding interaction constraint. Thus, the binding interaction constraint method is easy to implement, high in interaction precision, high in stability and higher in experience feeling in a virtual scene, more conforms to the virtual environment immersion characteristic, and solves thebinding interaction constraint problem of the human body mapping posture body and the virtual rotating body.

The embodiment of the invention discloses a determination method, device and system for a corresponding relation between a spatial position and a code of a handle. According to the method, the spatialposition of the handle and a grayscale image sent by a camera are acquired, and the handle emits infrared rays at preset flicker frequency f<handle>, wherein f<handle> is greater than f<camera>*K / 2C,f<handle> is the flicker frequency of the handle, f<camera> is refresh frequency of the camera, K is the number of scanning lines of a camera resolution, and C is the number of scanning lines occupied by the grayscale image of the handle; the grayscale image is obtained after the camera photographs the handle, and the grayscale image is in a black and white stripe shape; and the code of the handle under the spatial position is determined according to the grayscale image. Through the determination method, device and system, the situation that when a processor is disturbed by the color of a background article when recognizing the code of the handle according to a handle color collected by the camera can be effectively avoided, the corresponding relation between the spatial position and thecode of the handle can be accurately determined, and the interaction precision of a VR system is improved.

The invention requests to protect a tracking registration precision correction method in an augmented reality system. The method comprises the following specific operation processes: performing clustering analysis on spatial data; Locally employing t-S fuzzy system preprocessing; Deep learning is utilized globally for training; Preliminarily correcting a training target of the convolutional neuralnetwork by using a deep learning algorithm; Solving by adopting a least square support vector machine; And outputting a correction result when the precision reaches the application requirement, or else returning the data to continue deep learning training.

The invention discloses a robot, a method, a device and a medium, and the robot projects an advancing indication diagram in real time in the operation process, and obtains an obstacle area existing on a road surface in the operation process; detecting whether there is an overlapping region between the advancing indication diagram and the obstacle region; and when an overlapping region exists between the advancing indication diagram and the obstacle region, adjusting the advancing indication diagram according to the overlapping region, so that the overlapping region does not exist between the advancing indication diagram and the obstacle region. The advancing indication diagram is adjusted according to the curve overlapping area, so that the advancing indication diagram emitted by the robot can be dynamically deformed according to different obstacle areas, the adjusted advancing indication diagram is not overlapped with the obstacle areas, and then information interaction between the robot and the obstacle is achieved. And the information interaction efficiency and the information interaction accuracy between the robot and the person are improved.

The present disclosure discloses an active pen for a touch display screen, a touch display screen and a touch control input system, to improve human-machine interaction accuracy of operating the touch display screen with the active pen by a user. The active pen for the touch display screen comprises a transparent elastic pen point, an optical receiver and an excitation signal emitting apparatus, wherein the contact area between the transparent elastic pen point and the touch display screen is increased with an increase in an operating force, the optical receiver is configured to receive detection light which comes from the touch display screen and penetrates through the transparent elastic pen point, meanwhile, the excitation signal emitting apparatus is in signal connection with the optical receiver, so as to emit an excitation signal of a corresponding intensity to the touch display screen in accordance with the intensity of the detection light received by the optical receiver.

The invention belongs to the technical field of intelligent equipment, and particularly relates to an intelligent sensing glove and an intelligent sensing method. The intelligent sensing glove comprises a cover shell and a fastening belt. The fastening belt is arranged at the side end of the cover shell, and a sensor group is arranged on the inner side of the cover shell. The cover shell is fixed through the fastening belt, so that the sensor group on the inner side of the cover shell is in close contact with the skin of the hand back of a wearer. The action information of muscle tendons of the hand back is sensed by sensors in real time, and the motion information of the fingers of the wearer is identified according to the action information of the muscle tendons of the hand back. According to the intelligent sensing glove and the intelligent sensing method, the strip-shaped mechanical sensors are distributed on the inner side of the intelligent sensing glove at intervals, so that the number of the sensors is reduced greatly, and the reduction of cost and the reduction of the weight of the intelligent sensing glove are facilitated. The intelligent sensing glove is simple in structure and more convenient to use. The intelligent sensing glove and the intelligent sensing method can improve interaction accuracy, are strong in system work stability, and are not prone to being affected by environmental factors.

The invention relates to the technical field of artificial intelligence, and provides a cross-modal feature alignment method and device, equipment, a storage medium and a program product. The method comprises the following steps: inputting multi-modal interaction data into a feature extraction module of a multi-modal agent to obtain visual features, voice features and text features output by the feature extraction module; inputting the visual features, the voice features and the text features into a feature alignment module of a multi-modal agent to obtain a fusion feature vector output by the feature alignment module; and inputting the fusion feature vector into a semantic conflict detection and correction module of the multi-modal agent to obtain a correction feature output by the semantic conflict detection and correction module. Through the above mode, information loss caused by forced feature alignment can be avoided, the precision and robustness of feature alignment can be improved, the feature alignment effect can be optimized, the consistency and accuracy of aligned features can be ensured, and the interaction accuracy of the multi-modal intelligent agent can be improved.

The invention discloses a medical institution-oriented AI data governance method, which relates to the field of intelligent medical treatment, solves the problem that the existing AI data governance method has a poor medical data governance effect, and comprises the following steps: S1, acquiring medical text data and medical image data, dividing the medical text data into independent medical analysis texts, and storing the medical analysis texts in a database; s2, carrying out diagnosis matching analysis on the medical analysis image according to the medical institution acquisition data, obtaining an image diagnosis reasonability index corresponding to the medical analysis image according to the analysis result, and carrying out diagnosis matching analysis on the medical analysis image according to the analysis result; and S3, according to the medical image preliminary screening data and the data collected by the medical institution, medical data treatment is carried out to obtain treatment qualified data, and the AI data treatment effect can be improved.

The invention discloses a voice interaction method and device. The method includes: determining at least one piece of voice data collected by a plurality of microphones Mic; sending the at least one piece of voice data to a server; receiving each piece of identification sent by the server Information and its corresponding feedback results; perform semantic analysis on each piece of identification information to obtain a semantic recognition score for each piece of identification information; determine the feedback result corresponding to any piece of identification information with the highest semantic recognition score as output data; Output the output data. The embodiments of the present invention improve the accuracy of voice interaction.

The invention discloses a house relocation business management and collaboration platform, and relates to the technical field of house relocation management. The invention comprises a basic data management system, an account management system, a collaborative office system, a collection service system, a platform monitoring system, a data analysis system and a database. The basic data management system is mainly used for carrying out basic management on various information; the account management system is used for managing various accounts; the collaborative office system is mainly composed of an online message leaving module, a work collaborative module and a common question answering module. The collection service system comprises a service of a mobile terminal and a service of a computer terminal; the platform monitoring system comprises access log monitoring, user login monitoring and browsing footprint monitoring; the data analysis system comprises data analysis of a mobile terminal and data analysis of a computer terminal. Various problems existing in house relocation work are solved.

The embodiment of the invention provides a data forwarding control method and device, a storage medium and an electronic device. The method comprises the following steps: acquiring a data calling instruction; performing first processing on the data calling instruction to obtain a data detection instruction; sending the data detection instruction to a database; obtaining target data information fed back by the database based on the data detection instruction, and analyzing the target data information to obtain target data features of the target data information; and determining a target forwarding rule corresponding to the target data feature, and sending a target instruction to a database based on the target forwarding rule to instruct the database to perform data processing according to the target forwarding rule. According to the method and the device, the problem of data identification lag is solved, and the effect of improving the data interaction accuracy is further achieved.

The invention discloses a large screen interaction method based on a handheld device, belonging to the field of human-computer interaction. The method comprises the following steps of: (1) establishing a connection between a server and a client side, and configuring and sending the number of large screen grids to the server by the client side; (2) after a large screen is divided by the server, sending a code of a grid where a cursor is located currently to the client side, and initializing an interactive interface by the client side; (3) moving the client side, and sending an image collected in real time to the server; (4) estimating the displacement of the client side by the server, and mapping the displacement of the client side into motion of the cursor on the large screen; (5) after the client side is moved to ensure that the cursor of the large screen is located to a certain grid region, switching the interactive interface to the current grid region of the large screen by the client side, controlling the interactive interface of the client side and sending the input control information to the server; and (6) carrying out interaction operation by the server in the current grid range of the large screen according to the control information. The large screen interaction method has the characteristics of strong interactivity, low cost, easiness for operation and the like.

The invention discloses a laser cursor monitoring method and device and a storage medium. The laser cursor monitoring method comprises the steps: monitoring whether a light spot is projected to a display screen of the laser cursor monitoring device or not through a monitoring sensor in the display screen; when it is monitored that a light spot is projected to the display screen, obtaining the attribute information of the light spot; and monitoring whether a laser cursor exists in the display screen or not according to the attribute information of the light spot. According to the method, the interaction precision with the television can be improved.

The application discloses an intelligent ring and a ring interaction control method based on Morse code, and relates to the technical field of ring interaction control methods, and aims at solving the problem of single interaction instruction of the existing intelligent ring. The intelligent ring comprises a ring main body, an annular groove is arranged on the inner side surface of the ring main body, an annular circuit board capable of being wirelessly connected with a smart device is arranged in the annular groove, an induction assembly is arranged on the annular circuit board and is electrically connected with the annular circuit board and is in contact with the ring main body so that the ring main body is touched to input an instruction, an annular light-emitting strip is arranged on one side of the annular circuit board in the annular groove and is electrically connected with the annular circuit board, a light-transmitting area is arranged on the outer side surface of the ring main body and is opposite to the annular light-emitting strip, a ring inner ring is arranged on the cover of the annular groove so that the annular circuit board and the annular light-emitting strip are enclosed in the annular groove, Morse code encoding logic is combined with point pressing and long pressing operations of the intelligent ring, a large number of interaction instructions are generated through the combination of points, lines and intervals, the problem of single interaction instruction of the existing intelligent ring is solved, and the interaction expansibility is greatly improved.

The present invention relates to a device for interacting with vessel images, the device (14) comprising: an interface unit (22); a processing unit (20); wherein the interface unit (22) comprises: a display (21); and an input setup (23); wherein the display (21) is configured to display a vessel image (24); wherein the input setup (23) is configured to receive a user input in relation to the vessel image (24); wherein the processing unit (20) is configured to: determine, for at least one vessel (26) in the vessel image (24), a vessel contour (30); determine, from the user input, an identifier position (36) in the vessel image (24); indicate at least a portion (38) of the vessel contour (30) in the vessel image (24), if the determined identifier position (36) is spaced apart from the vessel contour (30) by a distance (37) within a predefined distance range; determine, from the user input, a drag direction (42); and move the indicated portion (38) along the vessel contour (30) based on the determined drag direction (42).

The embodiment of the invention provides a virtual article interaction method, a virtual article interaction device, a medium and equipment. The method comprises: in response to a first interaction triggering operation, determining at least one first virtual item as a target virtual item, wherein a game role is located in an interaction range corresponding to the first virtual item; in response to the second interaction triggering operation, obtaining a second virtual article in an interaction range corresponding to the first virtual article, and determining the second virtual article as a target virtual article; and in response to the third interaction triggering operation, executing an interaction operation with the target virtual article. Visibly, by implementing the technical scheme of the embodiment of the invention, the interaction operation efficiency can be improved.

One or more embodiments of this specification provide a method and device for verifying a graphic code, the method comprising: scanning a target graphic code to generate a scanning result for the target graphic code; , to determine whether the target graphic code meets the preset conditions; if so, determine that this scan code operation is an operation requesting to verify whether the target graphic code belongs to the applicant user; generate the first code indicating that the graphic code verification is successful and the target graphic code belongs to the applicant user - Interactive information. By generating targeted feedback information based on the verification result of the graphic code, the user can know that the feedback information is generated according to the verification requirements of the graphic code requested by the user, and it is not a simple error prompt. The feedback result of the authenticity verification of the graphic code is delivered to the user according to the user's graphic code verification requirements, which improves the accuracy of the interaction between the terminal and the user, thereby improving the user experience.