A touch recognition method, device and storage medium

Abstract

The application provides a touch recognition method and device and a storage medium, relates to the technical field of electronics, and aims to solve the technical problem that the prior art cannot accurately determine whether there is a touch operation on a screen. The touch recognition method comprises the following steps: in response to a touch operation performed by a user on the screen, acquiring capacitance current change information of the screen; and determining a touch mode of the touch operation according to the capacitance current change information of the screen and a graph neural network. The application can improve the efficiency of touch recognition.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronics, and in particular to a touch recognition method and device and a storage medium. BACKGROUND

[0002] With the continuous development of electronic technology, touch screens have been widely used in various electronic products, such as smart phones, smart home appliances, tablet computers, automatic teller machines, and vending machines. Touch screens support touch operations, and the touch screen technology has also developed rapidly. Touch screens have evolved from resistive touch screens to capacitive touch screens, and capacitive touch screens are currently the main type of touch screens used in various electronic products.

[0003] Capacitive touch screens work by using the current induction of the human body. In smart manufacturing or consumer user scenarios, when a user performs a touch operation on a capacitive touch screen, the user may operate the screen with water or oil, which can affect the click accuracy on the screen. Because water and oil can change the capacitance value of the screen, the screen may misjudge that a water droplet point is a finger operation point, or because the capacitance change of the water droplet is too large and the amount of electricity carried by the finger is not enough to change the capacitance range of the current screen local area (the capacitance change of the water droplet is too large), the screen cannot accurately determine that there is a finger click. SUMMARY

[0004] The present application provides a touch recognition method, device and storage medium, which solves the technical problem that general technology cannot meet the real-time interference suppression requirement.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0006] In a first aspect, a touch recognition method is provided, comprising:

[0007] In response to a touch operation performed by a user on a screen, obtaining capacitance current change information of the screen;

[0008] Determining a touch mode of the touch operation according to the capacitance current change information of the screen and a graph neural network.

[0009] Optionally, determining the touch mode of the touch operation according to the capacitance current change information of the screen and the graph neural network. The method comprises:

[0010] Determining the touch mode of the touch operation according to the capacitance current change information of the screen and a pre-constructed comprehensive graph neural topology structure; the touch mode includes any one of the following: a foreign object touch mode, a non-foreign object touch mode, and a superimposed touch mode of the foreign object touch mode and the non-foreign object touch mode; the comprehensive graph neural topology structure is constructed according to the graph neural network; and the comprehensive graph neural topology structure is used to represent the change relationship information of the capacitance current of each detection point in a plurality of detection points on the screen.

[0011] Optionally, the construction process of the integrated graph neural topology is as follows:

[0012] The graph neural topology construction process is performed for the touch operations of the multiple touch modes received by the screen to obtain multiple graph neural topology substructures corresponding to the multiple touch operations one by one;

[0013] The touch operations of the multiple touch modes include touch operations of a non-foreign object touch mode and touch operations of multiple foreign object touch modes with different foreign object contents; the multiple graph neural topology substructures include a non-foreign object graph neural topology and multiple foreign object graph neural topologies corresponding to the multiple different foreign object contents one by one;

[0014] The graph neural topology construction process includes determining the capacitance current change information of each detection point in the multiple detection points; determining the edge relationship weight between one detection point and other detection points according to the capacitance current change information of the one detection point to obtain multiple edge relationship weights corresponding to the multiple detection points one by one; determining multiple first matrix operators and second matrix operators corresponding to the multiple detection points according to the multiple edge relationship weights; and determining the graph neural topology substructure according to the multiple first matrix operators and the second matrix operators; the graph neural topology substructure Q 子结构 satisfies the following formula:

[0015] Q 子结构 =norm[sigmod{A 总 *B}];

[0016] wherein, sigmod is a preset function, Norm is a normalization function, A 总 is the sum of the multiple first matrix operators, and B is the second matrix operator;

[0017] The first matrix operator of one detection point is used to represent the adjacent relationship between the one detection point and other detection points; and the second matrix operator is used to represent the superimposed influence of the other detection points on the one detection point;

[0018] The integrated graph neural topology is determined according to the multiple graph neural topology substructures and convolution operations; and the integrated graph neural topology satisfies the following formula:

[0019] Q 综合 =Q 非异物 *Q 异物 ;

[0020] wherein, Q 综合 represents the integrated graph neural topology; Q 非异物 represents the non-foreign object graph neural topology; Q 异物 represents one foreign object graph neural topology; and * represents the convolution operation.

[0021] Optionally, the capacitance current change information of each detection point in the plurality of detection points is determined, including:

[0022] The screen is divided into an N*M horizontal and vertical array to obtain N*M nodes; N and M are positive integers; one node corresponds to one detection point;

[0023] A target node in the N*M nodes is determined; the target node includes: a center point of the N*M nodes, an upper vertex of a center axis of the N*M nodes, a left vertex of the N*M nodes, and a left vertex of the center axis of the N*M nodes;

[0024] The touch operation is performed on the target node to obtain the change relationship information of the capacitance current of the target node and the change relationship information of the capacitance current of other nodes.

[0025] Optionally, according to the capacitance current change information of one detection point, the edge relationship weight between the one detection point and other detection points is determined to obtain a plurality of edge relationship weights corresponding to the plurality of detection points, including:

[0026] According to the change relationship information of the capacitance current of the target node and the change relationship information of the capacitance current of other nodes, the edge relationship weight between the target node and other nodes is determined;

[0027] According to the edge relationship weight between the target node and other nodes and a preset algorithm, a plurality of edge relationship weights corresponding to the plurality of detection points are determined; the preset algorithm includes: a distance algorithm, and / or, a symmetry algorithm.

[0028] Optionally, according to the plurality of edge relationship weights, a plurality of first matrix operators and second matrix operators corresponding to the plurality of detection points are determined, including:

[0029] The first operation is performed on each detection point in the plurality of detection points to obtain a plurality of first matrix operators; the first operation includes: determining the sum of values in the edge relationship weight of one detection point as the first matrix operator of the one detection point;

[0030] The second matrix operator is determined; the value of the Nth row and the Mth column in the second matrix operator is: the sum of values corresponding to the edge relationship of the Nth row and the Mth column detection point in the edge relationship weight of other N*M-1 detection points except the Nth row and the Mth column detection point in the plurality of detection points.

[0031] Optionally, according to the capacitance current change information of the screen and a pre-constructed comprehensive graph neural topology structure, the touch mode of the touch operation is determined, including:

[0032] According to the capacitance current change information of the screen and the integrated graph neural topology structure, a first value G and a second value K are determined; the first value G is used to represent a capacitance change matrix; and the second value K is used to represent a current change matrix.

[0033] The first value G and the second value K satisfy the following formula:

[0034] G = relue < sigmod [Q 综合 *C]>;

[0035] K = relue < sigmod [Q 综合 *I]>;

[0036] Wherein, Q 综合 is used to represent the integrated graph neural topology structure; C is used to represent the capacitance change information of the screen; I is used to represent the current change information of the screen; relue is used to represent an activation function; and sigmoid is used to represent a sigmoid function.

[0037] When the first value G is greater than a preset value, and / or the second value K is greater than a preset value, it is determined that the touch mode is the superimposed touch mode.

[0038] When the first value G is less than or equal to a preset value, and the second value K is less than or equal to a preset value, it is determined that the touch mode is the foreign matter touch mode or the non-foreign matter touch mode.

[0039] Optionally, the touch recognition method further comprises:

[0040] When the touch mode of the touch operation is the superimposed touch mode, according to the capacitance current change information of the touch point corresponding to the touch operation, the foreign matter content of the touch operation is determined, and a target foreign matter graph neural topology structure corresponding to the foreign matter content is determined.

[0041] According to the non-foreign matter graph neural topology structure, the target foreign matter graph neural topology structure, the capacitance change matrix, the current change matrix and the edge relationship weight of the touch point, a first matrix X and a second matrix Y are determined; the first matrix X and the second matrix Y satisfy the following formula:

[0042] X = FFT -1 [FFT(Q 非异物 )*FFT(G)]*sum(W);

[0043] Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)]*sum(W);

[0044] Wherein, FFT -1 is an inverse Fourier transform function, FFT is a fast Fourier transform function; Q 目标异物W represents an edge relationship weight of the touch point;

[0045] When the first matrix X and the second matrix Y satisfy the target relationship, no capacitive compensation is performed on the touch point; the target relationship is:

[0046] (X2-X1) / X1>(Y2-Y1) / Y1;

[0047] X1 represents a value corresponding to the touch point in the first matrix X; X2 represents a value corresponding to other points in the first matrix X except the touch point; Y1 represents a value corresponding to the touch point in the second matrix Y; and Y2 represents a value corresponding to other points in the second matrix Y except the touch point;

[0048] When the first matrix X and the second matrix Y do not satisfy the target relationship, the reference capacitance of the touch point is adjusted until the touch points in the screen all satisfy the target relationship.

[0049] In a second aspect, a touch recognition device is provided, including: an acquisition unit and a processing unit.

[0050] The acquisition unit is configured to acquire capacitive current change information of the screen in response to a touch operation performed by a user on the screen.

[0051] The processing unit is configured to determine a touch mode of the touch operation according to the capacitive current change information of the screen and a graph neural network.

[0052] Optionally, the processing unit is specifically configured to:

[0053] determine the touch mode of the touch operation according to the capacitive current change information of the screen and a comprehensive graph neural topology structure constructed in advance; the touch mode includes any one of the following: a foreign matter touch mode, a non-foreign matter touch mode, and a superimposed touch mode of the foreign matter touch mode and the non-foreign matter touch mode; the comprehensive graph neural topology structure is constructed according to the graph neural network; and the comprehensive graph neural topology structure is used to represent a change relationship of a capacitive current of each detection point in a plurality of detection points on the screen.

[0054] Optionally, the processing unit is further configured to perform graph neural topology structure construction processing on a plurality of touch operations of the touch mode received by the screen, to obtain a plurality of graph neural topology substructures corresponding to the plurality of touch operations one by one.

[0055] The plurality of touch operations of the touch mode include: a touch operation of a non-foreign matter touch mode and touch operations of a plurality of foreign matter touch modes with different foreign matter contents; and the plurality of graph neural topology substructures include: a non-foreign matter graph neural topology structure and a plurality of foreign matter graph neural topology structures corresponding to the plurality of different foreign matter contents one by one.

[0056] The graph neural topology construction processing includes: determining, for each detection point in a plurality of detection points, capacitance current change information of the detection point; determining, according to the capacitance current change information of one detection point, an edge relationship weight between the one detection point and other detection points, to obtain a plurality of edge relationship weights corresponding to the plurality of detection points one by one; determining, according to the plurality of edge relationship weights, a plurality of first matrix operators and a plurality of second matrix operators corresponding to the plurality of detection points one by one; and determining, according to the plurality of first matrix operators and the plurality of second matrix operators, a graph neural topology substructure; the graph neural topology substructure Q 子结构 satisfies the following formula:

[0057] Q 子结构 = norm[sigmod{A 总 *B}];

[0058] wherein, sigmod is a preset function, Norm is a normalization function, A 总 is a sum of the plurality of first matrix operators, and B is the second matrix operator;

[0059] The first matrix operator of one detection point is used to represent an adjacent relationship between the one detection point and other detection points; and the second matrix operator is used to represent an additive effect of the other detection points on the one detection point.

[0060] The processing unit is further configured to determine, according to the plurality of graph neural topology substructures and a convolution operation, a comprehensive graph neural topology structure; and the comprehensive graph neural topology structure satisfies the following formula:

[0061] Q 综合 = Q 非异物 * Q 异物 ;

[0062] wherein, Q 综合 represents the comprehensive graph neural topology structure; Q 非异物 represents the non-foreign object graph neural topology structure; Q 异物 represents one foreign object graph neural topology structure; and * represents the convolution operation.

[0063] Optionally, the processing unit is specifically configured to:

[0064] divide the screen into an N*M horizontal and vertical array to obtain N*M nodes; N and M are positive integers; and one node corresponds to one detection point.

[0065] determine a target node in the N*M nodes; the target node includes: a center point of the N*M nodes, an upper top point of a center axis of the N*M nodes, a left top point of the N*M nodes, and a left top point of the center axis of the N*M nodes.

[0066] The touch operation is performed on the target node to obtain the change relationship information of the capacitive current of the target node and the change relationship information of the capacitive current of the other nodes.

[0067] Optionally, the processing unit is specifically configured to:

[0068] According to the change relationship information of the capacitive current of the target node and the change relationship information of the capacitive current of the other nodes, the edge relationship weight between the target node and the other nodes is determined.

[0069] According to the edge relationship weight between the target node and the other nodes and a preset algorithm, a plurality of edge relationship weights corresponding to a plurality of detection points are determined; the preset algorithm includes a distance algorithm and / or a symmetry algorithm.

[0070] Optionally, the processing unit is specifically configured to:

[0071] The first operation is performed on each detection point in the plurality of detection points to obtain a plurality of first matrix operators; the first operation includes determining the sum of the values in the edge relationship weight of one detection point as the first matrix operator of the detection point.

[0072] The second matrix operator is determined; the value of the Nth row and the Mth column in the second matrix operator is the sum of the values corresponding to the edge relationship of the Nth row and the Mth column detection point in the edge relationship weight of the other N×M-1 detection points except the Nth row and the Mth column detection point in the plurality of detection points.

[0073] Optionally, the processing unit is specifically configured to:

[0074] According to the capacitive current change information of the screen and the integrated graph neural topology structure, the first value G and the second value K are determined; the first value G is used to represent the capacitance change matrix; and the second value K is used to represent the current change matrix.

[0075] The first value G and the second value K satisfy the following formula:

[0076] G=relue<sigmod[Q 综合 *C]>;

[0077] K=relue<sigmod[Q 综合 *I]>;

[0078] Wherein, Q 综合 is used to represent the integrated graph neural topology structure; C is used to represent the capacitance change information of the screen; I is used to represent the current change information of the screen; relue is used to represent the activation function; and sigmoid is used to represent the sigmoid function.

[0079] determine that the touch mode is the superimposed touch mode when the first value G is greater than a preset value and / or the second value K is greater than a preset value;

[0080] determine that the touch mode is the foreign matter touch mode or the non-foreign matter touch mode when the first value G is less than or equal to a preset value and the second value K is less than or equal to a preset value.

[0081] Optionally, the processing unit is further configured to determine a foreign matter content of the touch operation according to the capacitive current change information of the touch point corresponding to the touch operation when the touch mode of the touch operation is the superimposed touch mode, and determine a target foreign matter graph neural topology structure corresponding to the foreign matter content.

[0082] The processing unit is further configured to determine a first matrix X and a second matrix Y according to the non-foreign matter graph neural topology structure, the target foreign matter graph neural topology structure, the capacitive change matrix, the current change matrix, and an edge relationship weight of the touch point; the first matrix X and the second matrix Y satisfy the following formula:

[0083] X = FFT -1 [FFT(Q 非异物 )*FFT(G)]*sum(W);

[0084] Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)]*sum(W);

[0085] wherein FFT -1 is an inverse Fourier transform function, FFT is a fast Fourier transform function; Q 目标异物 is used to represent the target foreign matter graph neural topology structure; W is used to represent the edge relationship weight of the touch point.

[0086] The processing unit is further configured to not perform capacitive compensation on the touch point when the first matrix X and the second matrix Y satisfy a target relationship; the target relationship is:

[0087] (X2-X1) / X1 > (Y2-Y1) / Y1.

[0088] wherein X1 is used to represent a value corresponding to the touch point in the first matrix X; X2 is used to represent a value corresponding to other points except the touch point in the first matrix X; Y1 is used to represent a value corresponding to the touch point in the second matrix Y; and Y2 is used to represent a value corresponding to other points except the touch point in the second matrix Y.

[0089] The processing unit is further configured to adjust a reference capacitance of the touch point until the touch points in the screen all satisfy the target relationship when the first matrix X and the second matrix Y do not satisfy the target relationship.

[0090] In a third aspect, a touch recognition apparatus is provided, comprising a memory and a processor; the memory is configured to store computer-executable instructions, and the processor is connected to the memory through a bus; when the touch recognition apparatus is running, the processor executes the computer-executable instructions stored in the memory, so that the touch recognition apparatus executes the touch recognition method in the first aspect.

[0091] The touch recognition apparatus can be a network device, or a part of the apparatus, such as a chip system in the network device. The chip system is configured to support the network device to implement the functions involved in the first aspect and any possible implementation manner thereof, such as obtaining, determining, and sending the data and / or information involved in the touch recognition method. The chip system comprises a chip, and can further comprise other discrete devices or circuit structures.

[0092] In a fourth aspect, a computer-readable storage medium is provided, which comprises computer-executable instructions. When the computer-executable instructions are running on a computer, the computer is caused to execute the touch recognition method in the first aspect.

[0093] In a fifth aspect, a computer program product is further provided, which comprises computer instructions. When the computer instructions are running on a touch recognition apparatus, the touch recognition apparatus is caused to execute the touch recognition method in the first aspect.

[0094] It should be noted that the computer instructions can be stored on the computer-readable storage medium in whole or in part. The computer-readable storage medium can be packaged together with the processor of the touch recognition apparatus, or packaged separately from the processor of the touch recognition apparatus, and the embodiments of the present application do not limit the same.

[0095] The second aspect, the third aspect, the fourth aspect, and the fifth aspect in the present application can refer to the detailed description of the first aspect.

[0096] In the embodiments of the present application, the name of the touch recognition apparatus does not limit the apparatus or the functional module itself, and these apparatuses or functional modules can appear in other names in actual implementation. For example, the receiving unit can also be referred to as a receiving module, a receiver, etc. As long as the functions of each apparatus or functional module are similar to those in the present application, they belong to the scope of the claims of the present application and equivalent technologies.

[0097] The technical solutions provided in the present application at least have the following beneficial effects:

[0098] Based on any of the above aspects, embodiments of this application provide a touch recognition method that can acquire information on changes in the capacitive current of the screen in response to a user's touch operation on the screen. Then, the touch mode of the touch operation can be determined based on the screen's capacitive current change information and a graph neural network. Thus, after acquiring the screen's capacitive current change information, this application can utilize a graph neural network to quickly and accurately determine the touch mode of the touch operation, improving the efficiency of touch recognition and solving the technical problem of existing technologies being unable to accurately determine whether there is a touch operation on the screen.

[0099] The beneficial effects of the first, second, third, fourth, and fifth aspects of this application can all be referred to in the analysis of the above-mentioned beneficial effects, and will not be repeated here. Attached Figure Description

[0100] Figure 1 This is a schematic diagram illustrating an application scenario of a touch recognition method provided in an embodiment of this application;

[0101] Figure 2 A schematic diagram of the hardware structure of the touch recognition device provided in the embodiments of this application. Figure 1 ;

[0102] Figure 3 A schematic diagram of the hardware structure of the touch recognition device provided in the embodiments of this application. Figure 2 ;

[0103] Figure 4 A flowchart illustrating a touch recognition method provided in this application embodiment. Figure 1 ;

[0104] Figure 1 Node connection diagram of a touch recognition method provided in this application embodiment Figure 6 ;

[0105] Figure 2 Node connection diagram of a touch recognition method provided in this application embodiment Figure 7 ;

[0106] Figure 1 A matrix illustration of a touch recognition method provided in this application embodiment. Figure 8 ;

[0107] Figure 2 A flowchart illustrating a touch recognition method provided in this application embodiment. Figure 9 ;

[0108] Figure 3 A flowchart illustrating a touch recognition method provided in this application embodiment. Figure 10 ;

[0109] Figure 2 A matrix diagram of a touch recognition method provided by an embodiment of the present application Figure 11

[0110] Figure 1 A structure diagram of a touch recognition device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0111] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0112] It should be noted that in the embodiments of the present application, the words such as “exemplary” or “for example” are used to represent an example, illustration or description. Any embodiment or design scheme described as “exemplary” or “for example” in the embodiments of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the words such as “exemplary” or “for example” are intended to present the related concept in a specific manner.

[0113] In order to clearly describe the technical solutions in the embodiments of the present application, in the embodiments of the present application, the words such as “first”, “second” are used to distinguish the same items or similar items with basically the same function and effect, and those skilled in the art can understand that the words such as “first”, “second” are not used to limit the quantity and execution order.

[0114] As described in the background, the capacitive touch screen works by using the current induction of the human body. To realize multi-point touch control, the capacitive screen relies on increasing the electrodes of mutual capacitance. Simply speaking, the screen is divided into blocks, and a set of mutual capacitance modules is arranged in each region to work independently. Therefore, the capacitive screen can independently detect the touch control of each region, and after processing, it simply realizes multi-point touch control. When a user touches the capacitive screen, due to the human body electric field, the user's finger and the working surface form a coupling capacitance. Because the working surface is connected with a high-frequency signal, the finger absorbs a very small current, which flows out from the electrodes on the four corners of the screen, and theoretically the current flowing through the four electrodes is proportional to the distance from the finger to the four corners. The controller calculates the position by precisely calculating the proportion of the four currents. However, if there are water droplets or oil droplets on the screen, the touch point will drift, and the accuracy will decrease.

[0115] ​In the intelligent manufacturing or consumer user scenario, when the user performs a touch operation on the capacitive touch screen, the user may operate the screen with water or oil, which will affect the click accuracy on the screen. Because water and oil will change the capacitance value of the screen, the screen will misjudge that the point with water droplets is the finger operation point, or because the capacitance change of the water droplets is too large and the electric quantity carried by the finger is not enough to change the capacitance range of the current screen local area (the capacitance change of the water droplets is too large), so that the presence of the finger click cannot be accurately determined.

[0116] In general technology, the capacitive screen usually perceives the pressure point through the pressure sensor, or perceives the touch point through the infrared sensor, or perceives the shadow through the optical sensor to infer the touch point. However, the above several ways will increase the screen thickness and response time delay, and the sensitivity of the screen may decrease.

[0117] For example, the infrared sensor, if there is a water droplet next to the position of the finger operation, the finger next to the real operation finger touches the water droplet but does not touch the screen, and the infrared sensing superimposed capacitive screen will misjudge that the finger next to the real operation finger also touches the screen.

[0118] For example, the pressure sensor superimposed capacitive screen, if the finger is too large, it may affect a small part of the area next to the touch point, and is also easy to cause misjudgment and reduce the screen recognition accuracy.

[0119] In view of the above problems, the capacitive current change information of the screen can be obtained in response to the touch operation performed by the user on the screen. Then, the touch mode of the touch operation can be determined according to the capacitive current change information of the screen and the graph neural network. In this way, after obtaining the capacitive current change information of the screen, the graph neural network can be used to quickly and accurately determine the touch mode of the touch operation, thereby improving the efficiency of touch recognition and solving the technical problem that the prior art cannot accurately determine whether there is a touch operation on the screen.

[0120] Figure 1 The application scenario of the embodiments of the present application is shown in FIG. 1. Figure 1 As shown in the figure, the scenario is when the user uses an electronic device with a capacitive touch screen (i.e. the above-mentioned screen), the user touches the capacitive touch screen.

[0121] Among them, the electronic device can be any electronic device with a capacitive touch screen, including but not limited to smart phones, smart home appliances, tablet computers, automatic teller machines, vending machines, etc.

[0122] For example, Figure 1 For example, the user uses a smart phone.

[0123] In some possible implementation manners, the electronic device includes a housing, a touch screen, and electronic elements. In the embodiments of the present application, the touch screen is a touch display screen with a display function, to display relevant interfaces or information of the electronic device, as a channel for a user to view and interact with the electronic device.

[0124] In the embodiments, the touch screen is a capacitive touch screen. The housing and the touch screen jointly form a receiving space, and the electronic elements are accommodated in the receiving space. The housing can provide protection for the electronic elements, so as to avoid the internal elements from being dislocated or damaged due to external force impact, thereby prolonging the service life of the electronic device. The electronic elements can include at least one of a receiver, a microphone, a camera module, a battery, a central processing unit, a sensor, and the like, to realize relevant functions of the electronic device.

[0125] The electronic device can further include a master control chip and a touch screen control chip, and the master control chip and the touch screen control chip can be arranged in the housing. The master control chip is integrated on a mainboard of the electronic device, and is used to realize data processing, calculation, and control functions of other electronic elements of the electronic device. In some embodiments, the master control chip can be understood as a central processing unit of the electronic device. The touch screen control chip is electrically connected between the master control chip and the touch screen. The touch screen control chip is used to receive a control signal of the master control chip, drive the touch screen to display specified image data according to the control signal, and further used to determine touch data of a touch event position when the touch screen receives the touch event, and transmit the touch data to the master control chip, so that the master control chip determines instruction information corresponding to the touch event according to the touch data.

[0126] In a possible application scenario, as shown in Figure 2 , the capacitive touch screen of the electronic device has water droplets, and the user can touch the capacitive touch screen at the position with the water droplets (i.e., an overlapped touch mode of the foreign matter touch mode and the non-foreign matter touch mode).

[0127] In another possible application scenario, when the user's finger has no foreign matter such as water droplets or oil droplets, the user can directly touch the capacitive touch screen with the finger (i.e., the non-foreign matter touch mode).

[0128] In another possible application scenario, the user can also directly drop water droplets or oil droplets on the capacitive touch screen (i.e., the foreign matter touch mode).

[0129] The basic hardware structure of the electronic device includes Figure 3 or Figure 2 elements included in the touch recognition device. Hereinafter, the touch recognition device shown in Figure 3 and Figure 2 is taken as an example to introduce the hardware structure of the electronic device.

[0130] As shown in Figure 2As shown in FIG. 1, a hardware structure schematic diagram of the touch recognition device provided by the embodiment of the present application is shown. The touch recognition device comprises a processor 21, a memory 22, a communication interface 23 and a bus 24. The processor 21, the memory 22 and the communication interface 23 can be connected through the bus 24.

[0131] The processor 21 is the control center of the touch recognition device, which can be one processor or a general term of multiple processing elements. For example, the processor 21 can be a general central processing unit (CPU), or other general processors, etc. The general processor can be a microprocessor or any conventional processor, etc.

[0132] As an embodiment, the processor 21 can comprise one or more CPUs, such as the CPU 0 and the CPU 1 shown in FIG. 1. Figure 2

[0133] The memory 22 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable read-only memory (EEPROM), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and can be accessed by a computer, but is not limited to this.

[0134] In a possible implementation, the memory 22 can exist independently of the processor 21. The memory 22 can be connected to the processor 21 through the bus 24, and is used to store instructions or program codes. When the processor 21 invokes and executes the instructions or program codes stored in the memory 22, the touch recognition method provided by the embodiments of the present application can be implemented.

[0135] In the embodiments of the present application, the software programs stored in the memory 22 of the electronic device are different, so the functions implemented by the electronic device are different. The functions performed by each device will be described in combination with the flowcharts below.

[0136] In another possible implementation, the memory 22 can also be integrated with the processor 21.

[0137] ​The communication interface 23 is configured to connect the touch recognition device to other devices through a communication network, which can be an Ethernet, a wireless access network, a wireless local area network (WLAN), or the like. The communication interface 23 can include a receiving unit configured to receive data, and a sending unit configured to send data.

[0138] The bus 24 can be an industry standard architecture (ISA) bus, a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, or the like. The bus can be divided into an address bus, a data bus, a control bus, or the like. For the sake of representation, Figure 3 Only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus.

[0139] Figure 3 Another hardware structure of the touch recognition device in the embodiments of the present application is shown. As shown in the figure, Figure 2 The touch recognition device can include a processor 31 and a communication interface 32. The processor 31 is coupled to the communication interface 32.

[0140] The functions of the processor 31 can refer to the description of the processor 21 above. In addition, the processor 31 also has a storage function, which can function as the memory 22 described above.

[0141] The communication interface 32 is configured to provide data for the processor 31. The communication interface 32 can be an internal interface of the touch recognition device, or an external interface of the touch recognition device (equivalent to the communication interface 23).

[0142] It should be noted that, Figure 3 The structure shown in (or Figure 2 ) does not constitute a limitation on the touch recognition device. In addition to the components shown in (or Figure 3 (or Figure 4 ), the touch recognition device can include more or fewer components than shown, or combine certain components, or different component arrangements.

[0143] The touch recognition method provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0144] The touch recognition method provided by the embodiments of the present application comprises: a flow of constructing a comprehensive graph neural topology structure (referred to as a "comprehensive graph neural topology structure construction flow") by a touch recognition device, a flow of determining a touch mode of a touch operation according to the comprehensive graph neural topology structure (referred to as a "touch recognition flow") by the touch recognition device, and a flow of determining a capacitance compensation strategy of a touch point according to the comprehensive graph neural topology structure (referred to as a "capacitance compensation strategy determination flow") by the touch recognition device.

[0145] First, the "comprehensive graph neural topology structure construction flow" will be described.

[0146] As shown in Figure 5 , the specific method of the "comprehensive graph neural topology structure construction flow" comprises:

[0147] S401, the touch recognition device performs graph neural topology structure construction processing on the touch operations of multiple touch modes received by the screen to obtain multiple graph neural topology substructures corresponding to the multiple touch operations one by one.

[0148] Among them, the touch operations of multiple touch modes include: touch operations of non-foreign object touch modes, touch operations of foreign object touch modes with multiple different foreign object contents. The multiple graph neural topology substructures include: a non-foreign object graph neural topology structure and multiple foreign object graph neural topology structures corresponding to the multiple different foreign object contents one by one.

[0149] Specifically, the touch operation of the non-foreign object touch mode can be the touch operation of a finger or a touch pen on the screen without foreign objects. For ease of understanding, the present application takes the touch operation of the non-foreign object touch mode as an example to illustrate the touch operation of a finger on the screen.

[0150] The touch operation of the non-foreign object touch mode can be the touch operation of a finger or a touch pen on the screen without foreign objects. For ease of understanding, the present application takes the touch operation of the non-foreign object touch mode as an example to illustrate the touch operation of a finger on the screen.

[0151] The touch operation of the foreign object touch mode can be the touch operation of a foreign object (such as an oil droplet, a water droplet, etc.) on the screen. For ease of understanding, the present application takes the touch operation of the foreign object touch mode as an example to illustrate the touch operation of a water droplet on the screen.

[0152] The touch operation of the foreign object touch mode with multiple different foreign object contents can be the touch operation of a water droplet or an oil droplet with different milliliters (ML) on the screen.

[0153] For example, the touch operation of multiple foreign object touch modes with different foreign object contents can be a touch operation of a 1ML water droplet or oil droplet on the screen, a touch operation of a 1.2ML water droplet or oil droplet on the screen, a touch operation of a 1.3ML water droplet or oil droplet on the screen, etc.

[0154] In one feasible approach, the graph neural topology construction process includes:

[0155] S411, The touch recognition device determines the capacitance current change information of each of the multiple detection points.

[0156] In one feasible approach, the method by which a touch recognition device determines the capacitance current change information at each of multiple detection points specifically includes:

[0157] S411-1 The touch recognition device divides the screen into an N×M horizontal and vertical array to obtain N×M nodes.

[0158] Where N and M are positive integers. Each node corresponds to one detection point. Among the N×M nodes, the edges between each node represent the influence of the current node on its surrounding nodes.

[0159] For example, such as Figure 6 As shown, the touch recognition device can divide the screen into a 3×3 array, with the location of the finger touch point as a node, and the effect of the finger touch point on the capacitance changes of other surrounding locations, such as the connections between nodes, as edge relationships between nodes. For each touch point, the graph neural network topology is an M×N matrix array.

[0160] In one feasible approach, to avoid the impact of temperature and humidity on screen sensitivity, when the user first turns on the touch screen of the electronic device to which the touch recognition device belongs (i.e., the screen) (ensuring there are no foreign objects, such as water droplets, oil droplets, etc.), the touch recognition device can obtain the current environment based on the humidity H and temperature T sensors of the electronic device.

[0161] Next, the touch recognition device can divide the screen into an N×M horizontal and vertical array based on the ambient humidity and temperature.

[0162] The higher the humidity and the lower the temperature, the denser the array, meaning the larger the N and M numbers. Conversely, the lower the humidity and the higher the temperature, the sparser the array, meaning the smaller the N and M numbers.

[0163] In this case, the values ​​of N and M can satisfy the following formula:

[0164] M = a × 10 × [1 + lg(H / T)];

[0165] N = b × 10 × e (H / T) .

[0166] Where a, b, and e are constants.

[0167] In another possible implementation, the touch recognition device can divide the screen into an N×M horizontal and vertical array according to user needs, to obtain N×M nodes.

[0168] For example, when the screen is for a game application, it can be divided into a relatively dense array, i.e., with a large number of N and M. Conversely, when the screen is for a novel application, it can be divided into a relatively sparse array, i.e., with a small number of N and M.

[0169] S411-2, The touch recognition device determines the target node among N×M nodes.

[0170] The target nodes include: the center point of N×M nodes, the top vertex of the central axis of N×M nodes, the left vertex of N×M nodes, and the left vertex of the central axis of N×M nodes.

[0171] For example, such as Figure 6 As shown in (a) in the figure, Figure 6 The center point, the top vertex, the left vertex, and the left vertex of the central axis shown in (a) are the target node. If the touch point P is at the center of the screen, then N = M = 3, meaning P should correspond to the second row and second column of the matrix.

[0172] S411-3 The touch recognition device performs touch operations on all target nodes to obtain information on the change relationship of the capacitive current of the target node, as well as information on the change relationship of the capacitive current of other nodes.

[0173] Users can be at the central point ( Figure 6 The touch point P shown in (a) is clicked. During the screen click process, when the finger touches a certain point on the screen, the capacitance and current of that point and other points on the screen will change. The original touch screen module will accurately determine the position P of the touch point, as well as the capacitance C and current I changes corresponding to all position coordinates of the entire touch screen, and report them to the touch recognition device.

[0174] For example, such as Figure 6 As shown in (b), the capacitance values ​​of each node are: 0.2, -2.5, -0.5, -2.6, -5, -2.2, -0.1, -3, -0.1.

[0175] Touch recognition devices can construct graph neural network topologies and train the correlation between touch point capacitance and capacitance and current changes at surrounding coordinate points.

[0176] S412. The touch recognition device determines the edge relationship weights between a detection point and other detection points based on the capacitance current change information of a detection point, so as to obtain multiple edge relationship weights that correspond one-to-one with multiple detection points.

[0177] In one feasible approach, the touch recognition device determines the edge relationship weights between a detection point and other detection points based on the capacitance current change information of a detection point, thereby obtaining multiple edge relationship weights corresponding one-to-one with multiple detection points. Specifically, this method includes:

[0178] S412-1. The touch recognition device determines the edge relationship weight between the target node and other nodes based on the change relationship information of the capacitive current of the target node and the change relationship information of the capacitive current of other nodes.

[0179] For example, in combination Figure 6 As shown in (b), the touch recognition device can construct the edge relationship weight W between node P and other nodes based on the change in screen capacitance after touch.

[0180] When the finger touches node P, if the capacitance corresponding to the other N×M-1 node positions changes, it indicates that edge relationships should be established between these other positions and node P. In this case, the touch recognition device can be trained to identify the relationships between node P and the other N×M-1 nodes (…). Figure 6 (c) shows the set of edge relationship weights W between node P and 8 other nodes.

[0181] The weight set W includes N×M-1 w values.

[0182] If the capacitance changes significantly at other node locations, then the weight value w representing the edge relationship between that node and point P in the set should be large, for example, 0.99.

[0183] Correspondingly, if the change in capacitance at other node positions is less than the threshold 1, then the edge relationship weight w between point P and that position can be set to 0.001.

[0184] For example, W33 can be approximated as 0.02 (because for the bottom right corner 0.02×5=0.1), and W32 can be set as 0.6 (because for the midpoint of the bottom edge 0.6×5=3).

[0185] It should be noted that the specific algorithm for W mentioned above can refer to existing methods for calculating edge relation weights, and will not be elaborated here.

[0186] Next, the user can perform touch operations on other target nodes. The touch recognition device can obtain the edge relationship weight matrix between the other target nodes and the other N×M-1 nodes on the screen according to the above method.

[0187] S412-2, the touch recognition device determines a plurality of edge relationship weights corresponding to the plurality of detection points according to the edge relationship weights between the target node and other nodes and a preset algorithm.

[0188] The preset algorithm includes a distance algorithm and / or a symmetry algorithm.

[0189] In an implementable manner, when the preset algorithm is the distance algorithm, the touch recognition device can determine the distances between the target node and other nodes, and determine the edge relationship weights of the other nodes according to the distance algorithm.

[0190] In some embodiments, the bottom hardware chip of the mobile phone screen can report the distances between the target node and other nodes. For example, when a user touches the mobile phone screen, the chip below the mobile phone screen can sense and calculate the position touched by the user.

[0191] For example, when the target nodes are P1, P2, P3 and P4 respectively, a node Px in MxN is located between the four target nodes P1, P2, P3 and P4, and is at a distance D1 from the target node P1, a distance D2 from the target node P2, a distance D3 from the target node P3 and a distance D4 from the target node P4.

[0192] In this case, the edge relationship weight Wx of the node Px satisfies the following formula:

[0193] Wx=(W1xD1+W2xD2+W3xD3+W4xD4) / e(D1xD2xD3xD4).

[0194] W1 is the edge relationship weight of the target node P1, W2 is the edge relationship weight of the target node P2, W3 is the edge relationship weight of the target node P3, and W4 is the edge relationship weight of the target node P4.

[0195] In another implementable manner, when the preset algorithm is the symmetry algorithm, since the four target nodes selected are in the upper left part of the entire screen MxN matrix, the touch recognition device can determine the weight matrices W1, W2, …, Wmxn corresponding to the lower left corner, the lower right corner, the upper right corner and all other MxN position points according to the geometric symmetry of the screen.

[0196] In this way, the present application can infer the matrix W of all nodes in the full screen with only a few target nodes, saving time and resources, and the user does not need to repeatedly experiment.

[0197] S413, the touch recognition device determines a plurality of first matrix operators and a plurality of second matrix operators corresponding to the plurality of detection points according to the plurality of edge relationship weights.

[0198] The first matrix operator of one detection point is used to represent the adjacency relationship between the one detection point and other detection points. The second matrix operator is used to represent the superposition effect of other detection points on one detection point.

[0199] In an implementable manner, the touch recognition device determines the first matrix operator and the second matrix operator corresponding to the plurality of detection points according to the plurality of edge relationship weights, comprising:

[0200] S413-1, the touch recognition device performs a first operation on each of the plurality of detection points to obtain a plurality of first matrix operators.

[0201] The first operation comprises: determining the sum of the values in the edge relationship weight of one detection point as the first matrix operator of the one detection point.

[0202] For example, Figure 7 It can be seen that for the first matrix operator A of the P1 point, the P1 point is adjacent to P2, P3, and P4, so the first matrix operator A of the P1 point corresponds to the sum of all numbers in the first row and the first column (6.1+2.1+0.2+1.9+0.8+0.1+0.7+0.2+0.1) = 12.2.

[0203] The first row and the second column are all numbers in W2 added (1.2+5.5+1.2+0.7+1.4+0.8+0.2+0.1+0.3) = 11.4.

[0204] The first row and the third column are 0 because the P1 is not connected to the P5.

[0205] The second row and the first column are all numbers in W3 added (1.9+0.8+0.1+6.3+2.1+0.2+1.2+0.7+0.1) = 13.4.

[0206] The second row and the second column are all numbers in W4 added (0.2+2.5+0.5+2.6+5+2.2+0.1+3+0.1) = 16.2.

[0207] The second row and the third column and the third row and all columns are 0 because they are not adjacent to the P1.

[0208] For example, the first matrix operator A of the P4 point: the P4 is adjacent to the other eight nodes, so the first row and the first column of the A matrix of the P4 are all numbers in W1 added, the first row and the second column are all numbers in W2 added, and the third row and the third column are all numbers in W9 added.

[0209] Finally, the A matrix of the entire screen can be represented as Atotal. Atotal satisfies the following formula:

[0210] A总 =norm[i*(A1+A2……+Am×n)];

[0211] Where i is a preset parameter and Norm is a normalization function.

[0212] S413-2, The touch recognition device determines the second matrix operator.

[0213] The value in the Nth row and Mth column of the second matrix operator is the sum of the edge relation weights of the other N×M-1 detection points (excluding the detection point in the Nth row and Mth column) that have an edge relation with the detection point in the Nth row and Mth column.

[0214] Specifically, the second matrix operator B represents the cumulative effect of the other N×M-1 points on a single node. That is, the value in the Nth row and Mth column of the second matrix operator B is obtained by adding the values ​​of the weights W of the other N*M-1 nodes (excluding the point in the Nth row and Mth column on the screen) that have edge relationships with the corresponding point position.

[0215] For example, such as Figure 8 As shown, the value in the first row and first column of the second matrix operator B is obtained by adding the values ​​in the first row and first column of each of the eight W matrices corresponding to P2-P9:

[0216] 1.2 + 0.2 + 1.9 + 0.2 + 0.1 + 0.2 + 0.1 + 0.1 = 4.

[0217] The value in the first row and second column of the second matrix operator B is obtained by adding the values ​​in the first row and second column of each of the eight W matrices corresponding to P1, P3-P9:

[0218] 2.1 + 1.1 + 0.8 + 2.5 + 0.5 + 0.2 + 0.1 + 0.5 = 7.8.

[0219] The value in the first row and third column of the second matrix operator B is obtained by adding the values ​​in the first row and second column of each of the eight W matrices corresponding to P1-P2 and P4-P9: 0.2+1.2+0.1+0.5++0.1+0.2+0.1=2.4.

[0220] The value in the third row and second column of the second matrix operator B is obtained by adding the values ​​in the third row and second column of each of the eight W matrices corresponding to P1-P7 and P9: 0.2+0.1+0.1+0.7+3+0.5+1.7+2.5=8.8.

[0221] The value of the Nth row and Mth column in the second matrix operator B is calculated sequentially using the method described above, thus obtaining the second matrix operator B.

[0222] S414, the touch recognition device determines the graph neural topology substructure according to the plurality of first matrix operators and the second matrix operator.

[0223] The graph neural topology substructure Q 子结构 satisfies the following formula:

[0224] Q 子结构 = norm[sigmod{A 总 *B}];

[0225] Wherein, sigmod is a preset function, Norm is a normalization function, A 总 is the sum of the plurality of first matrix operators, and B is the second matrix operator.

[0226] Specifically, the general technology usually directly detects the capacitance C and current I of a certain node on the screen to infer the position of the touch point, but does not consider the capacitance and power changes at other positions on the screen. In this application, the touch recognition device can consider the capacitance and current change law of the entire screen, and report the capacitance and current change of all M*N nodes on the entire screen to the new management module. The new management module determines which node in M*N corresponds to the point, and the convolution operation of A matrix, B matrix and W can infer and predict the capacitance and current changes of the entire screen.

[0227] S402, the touch recognition device determines the comprehensive graph neural topology structure according to a plurality of graph neural topology substructures and convolution operations.

[0228] The above comprehensive graph neural topology structure satisfies the following formula:

[0229] Q 综合 = Q 非异物 *Q 异物 .

[0230] Wherein, Q 综合 is used to represent the comprehensive graph neural topology structure. Q 非异物 is used to represent the non-foreign object graph neural topology structure. Q 异物 is used to represent a foreign object graph neural topology structure. * is used to represent convolution operation.

[0231] Specifically, the touch recognition device can perform graph convolution operation on the non-foreign object graph neural topology structure and the one foreign object graph neural topology structure. In this way, the graph neural topology structure of the superimposed touch mode (such as the operation mode of finger superimposed water droplets) on the screen can be obtained, that is, the comprehensive graph neural topology structure, without repeatedly testing the scene of finger superimposed water droplets on the entire screen, thereby improving the efficiency of touch recognition.

[0232] Secondly, the integrated graph neural topology also includes NXM nodes, and the specific mathematical expression is also an NXM matrix. With the matrix of the integrated graph neural topology, as long as the capacitance and the electric quantity corresponding to all points (nodes) of the screen at a certain moment are input, it can be inferred whether the state has a water plus finger superposition condition.

[0233] Next, the "touch recognition process" will be described.

[0234] As shown in Figure 9 the specific method of the "touch recognition process" includes:

[0235] S801, the touch recognition device acquires the capacitance current change information of the screen in response to the touch operation performed by the user on the screen.

[0236] Specifically, when the user performs a touch operation on the screen, the touch screen module forms an NXM matrix C, I of the capacitance change and the current change of all points on the screen and reports it to the touch recognition device in real time.

[0237] S802, the touch recognition device determines the touch mode of the touch operation according to the capacitance current change information of the screen and the graph neural network.

[0238] In an implementable manner, the method for determining the touch mode of the touch operation according to the capacitance current change information of the screen and the graph neural network by the touch recognition device specifically includes:

[0239] The touch recognition device determines the touch mode of the touch operation according to the capacitance current change information of the screen and the pre-constructed integrated graph neural topology.

[0240] Among them, the touch mode includes any one of the following: foreign object touch mode, non-foreign object touch mode, superposition touch mode of foreign object touch mode and non-foreign object touch mode. The integrated graph neural topology is constructed according to the graph neural network. The integrated graph neural topology is used to represent the change relationship information of the capacitance current of each detection point in the plurality of detection points on the screen.

[0241] In an implementable manner, the method for determining the touch mode of the touch operation according to the capacitance current change information of the screen and the pre-constructed integrated graph neural topology by the touch recognition device specifically includes:

[0242] S821, the touch recognition device determines the first value G and the second value K according to the capacitance current change information of the screen and the integrated graph neural topology.

[0243] Among them, the first value G is used to represent the capacitance change matrix; the second value K is used to represent the current change matrix.

[0244] The first value G and the second value K satisfy the following formula:

[0245] G = relue < sigmod [Q 综合 * ].

[0246] K = relue < sigmod [Q 综合 * ].

[0247] Wherein, Q 综合 is used to represent the integrated graph neural topology structure. C is used to represent the capacitance change information of the screen. I is used to represent the current change information of the screen. relue is used to represent the activation function. Sigmoid is used to represent the sigmoid function.

[0248] Since the integrated graph neural topology structure includes all the relationship information of the N x M nodes on the screen, that is, the capacitance change relationship information between the nodes, if the capacitance current of a node changes, the capacitance current changes of other positions can be predicted synchronously, therefore, the above formula can be used to predict whether there is a water drop on the screen.

[0249] S822, when the first value G is greater than a preset value, and / or, the second value K is greater than a preset value, the touch recognition device determines that the touch mode is a superimposed touch mode.

[0250] In this case, the touch recognition device can determine the capacitance compensation strategy of the touch point according to the capacitance current change information of the touch point corresponding to the touch operation and the integrated graph neural topology structure.

[0251] S823, when the first value G is less than or equal to a preset value, and the second value K is less than or equal to a preset value, the touch recognition device determines that the touch mode is a foreign matter touch mode or a non-foreign matter touch mode.

[0252] In this case, the touch recognition device does not need to perform capacitance compensation on the touch point.

[0253] Next, the "capacitance compensation strategy determination process" will be described.

[0254] As shown in Figure 10 , the specific method of the above "capacitance compensation strategy determination process" includes:

[0255] S901, when the touch mode of the touch operation is a superimposed touch mode, the touch recognition device determines the foreign matter content of the touch operation according to the capacitance current change information of the touch point corresponding to the touch operation, and determines the target foreign matter graph neural topology structure corresponding to the foreign matter content.

[0256] Specifically, after determining the plurality of foreign matter graph neural topologies corresponding to the plurality of different foreign matter contents, the touch recognition device can determine which points on the screen dripped water according to the overall screen capacitance change.

[0257] Because the capacitance value of the water-dripped point decreases to a negative number and the capacitance value of the finger touch increases to a positive number, the touch recognition device can determine where water dripped and how much water dripped according to the capacitance change value, thereby determining the target foreign matter graph neural topology corresponding to the foreign matter content.

[0258] S902, the touch recognition device determines the first matrix X and the second matrix Y according to the non-foreign matter graph neural topology, the target foreign matter graph neural topology, the capacitance change matrix, the current change matrix, and the edge relationship weight of the touch point.

[0259] The first matrix X and the second matrix Y satisfy the following formula:

[0260] X = FFT -1 [FFT(Q 非异物 )*FFT(G)]*sum(W);

[0261] Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)]*sum(W)。

[0262] FFT -1 is an inverse Fourier transform function, and FFT is a fast Fourier transform function; Q 目标异物 is used to represent the target foreign matter graph neural topology; and W is used to represent the edge relationship weight of the touch point.

[0263] Specifically, the first matrix X is used to represent the capacitance on the nodes of all positions of the current screen respectively operating with the graph neural topology of only having a finger (i.e., Q 非异物 ), the graph neural topology of only having water droplets (i.e., Q 目标异物 ), and comparing which one has a higher influence on the entire screen.

[0264] S903, when the first matrix X and the second matrix Y satisfy the target relationship, the touch recognition device does not perform capacitance compensation on the touch point.

[0265] The target relationship is:

[0266] (X2-X1) / X1>(Y2-Y1) / Y1.

[0267] Wherein, X1 is used to represent the value corresponding to the touch point in the first matrix X. X2 is used to represent the value corresponding to the point other than the touch point in the first matrix X. Y1 is used to represent the value corresponding to the touch point in the second matrix Y. Y2 is used to represent the value corresponding to the point other than the touch point in the second matrix Y.

[0268] S904, when the first matrix X and the second matrix Y do not satisfy the target relationship, the touch recognition device adjusts the reference capacitance of the touch point until the touch point in the screen satisfies the target relationship.

[0269] Suppose the touch point is in the xth row and yth column, the value of the point in the xth row and yth column in the first matrix X and the second matrix Y can be determined.

[0270] For example, as shown in Figure 11 X1=5, Y1=8, then subtract the absolute value of the other MxN-1 points in the first matrix X and the second matrix Y from X1=5 and Y1=8, respectively, to obtain MxN-1 difference values of the two matrices, and compare the absolute values of the other MxN-1 difference values.

[0271] Specifically, when the first matrix X and the second matrix Y satisfy the target relationship, it means that the influence of the finger is greater than the influence of the water droplet at this position, that is, when the user operates at this position, the touch screen will recognize that it is a finger, and the point with water droplets will not affect this point, and there is no need to compensate the capacitance of the touch point.

[0272] When the first matrix X and the second matrix Y do not satisfy the target relationship, it means that the water droplet has a greater influence on this position than the finger, and if the finger touches this point, the screen may misjudge that it is a water droplet instead of a finger, and the reference capacitance of the touch point needs to be adjusted until the first matrix X and the second matrix Y satisfy the target relationship, that is, the influence of the water droplet on the touch point is no longer lower than the influence of the finger on the touch point, so as to ensure that when the user's finger touches the touch point, the screen can perceive that it is a finger instead of a water droplet.

[0273] The method for adjusting the reference capacitance of the touch point by the touch recognition device specifically comprises:

[0274] The touch recognition device selects the position point of (X-X1) / X1<(Y-Y1) / Y1, that is, the point whose influence of the water droplet is greater than the influence of the finger, wherein the position point with the largest difference is the point PX most affected by the water droplet.

[0275] The touch recognition device increases the capacitance Cx of these position points on the screen by a certain percentage. The specific increase ratio can be calculated according to the following formula:

[0276] h×[Cx-Cm];

[0277] Wherein h is an adjustment coefficient, Cx is the capacitance of the position point affected by the water droplet, and Cm is the average value of the capacitance of other position points around the previous moment.

[0278] Then, the touch recognition device repeats the above calculation to determine the point affected by the water droplet the least (because the capacitance of other points may change after the point affected by the water droplet the most is compensated), and then compensates until all touch points in the screen meet the target relationship.

[0279] The above mainly introduces the scheme provided by the embodiments of the application from the method aspect. To implement the above functions, it contains the hardware structure and / or software module corresponding to the execution of each function. Those skilled in the art should easily realize that, in combination with the units and algorithm steps of the examples described in the embodiments disclosed herein, the application can be implemented in the form of hardware or the combination of hardware and computer software. Whether a certain function is implemented in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the application.

[0280] The embodiments of the application can divide the function modules of the touch recognition device according to the above method examples. For example, each function module can be divided according to each function, or two or more functions can be integrated in one processing module. The integrated module can be implemented in the form of hardware or software function module. Optionally, the division of the modules in the embodiments of the application is illustrative, and is only a logical function division. In actual implementation, there can be another division manner.

[0281] As shown in Figures 4-9 , it is a structure schematic diagram of a touch recognition device provided by the embodiments of the application. The touch recognition device can be used to execute the touch recognition method shown in Figure 11 . ​ The touch recognition device shown in includes an acquisition unit 1001 and a processing unit 1002.

[0282] The acquisition unit 1001 is configured to acquire the capacitance current change information of the screen in response to the touch operation performed by the user on the screen.

[0283] The processing unit 1002 is configured to determine the touch mode of the touch operation according to the capacitance current change information of the screen and the graph neural network.

[0284] Optionally, the processing unit 1002 is specifically configured to:

[0285] Determine the touch mode of the touch operation according to the capacitive current change information of the screen and the pre-constructed integrated graph neural topology structure; the touch mode includes any one of the following: foreign object touch mode, non-foreign object touch mode, and superimposed touch mode of the foreign object touch mode and the non-foreign object touch mode; the integrated graph neural topology structure is constructed according to a graph neural network; and the integrated graph neural topology structure is used to represent the change relationship information of the capacitive current of each detection point in the plurality of detection points on the screen.

[0286] Optionally, the processing unit 1002 is further configured to perform graph neural topology structure construction processing on the plurality of touch operations of the plurality of touch modes received by the screen, to obtain a plurality of graph neural topology substructures corresponding to the plurality of touch operations one by one.

[0287] The plurality of touch operations of the plurality of touch modes include: touch operations of the non-foreign object touch mode and touch operations of the foreign object touch mode with a plurality of different foreign object contents; and the plurality of graph neural topology substructures include: a non-foreign object graph neural topology structure and a plurality of foreign object graph neural topology structures corresponding to the plurality of different foreign object contents one by one.

[0288] The graph neural topology structure construction processing includes: determining the capacitive current change information of each detection point in the plurality of detection points; determining the edge relationship weight between one detection point and other detection points according to the capacitive current change information of the one detection point, to obtain a plurality of edge relationship weights corresponding to the plurality of detection points one by one; determining a plurality of first matrix operators and a second matrix operator corresponding to the plurality of detection points according to the plurality of edge relationship weights; and determining the graph neural topology substructure according to the plurality of first matrix operators and the second matrix operator; the graph neural topology substructure Q 子结构 satisfies the following formula:

[0289] Q 子结构 = norm[sigmod{A 总 *B}];

[0290] Wherein, sigmod is a preset function, Norm is a normalization function, A 总 is the sum of the plurality of first matrix operators, and B is the second matrix operator;

[0291] The first matrix operator of one detection point is used to represent the adjacent relationship between the one detection point and other detection points; and the second matrix operator is used to represent the superimposed influence of the other detection points on the one detection point.

[0292] The processing unit 1002 is further configured to determine the integrated graph neural topology structure according to the plurality of graph neural topology substructures and convolution operation; and the integrated graph neural topology structure satisfies the following formula:

[0293] Q 综合 = Q 非异物 * Q 异物;

[0294] wherein Q 综合 for representing a comprehensive graph neural topology; Q 非异物 for representing a non-foreign matter graph neural topology; Q 异物 for representing a foreign matter graph neural topology; * for representing a convolution operation.

[0295] Optionally, the processing unit 1002 is specifically configured to:

[0296] divide the screen into an N*M horizontal and vertical array to obtain N*M nodes; N and M are positive integers; one node corresponds to one detection point;

[0297] determine a target node in the N*M nodes; the target node includes: a center point of the N*M nodes, an upper vertex of a center axis of the N*M nodes, a left vertex of the N*M nodes, and a left vertex of the center axis of the N*M nodes;

[0298] perform a touch operation on the target node to obtain change relationship information of a capacitive current of the target node and change relationship information of a capacitive current of other nodes.

[0299] Optionally, the processing unit 1002 is specifically configured to:

[0300] determine an edge relationship weight between the target node and other nodes according to the change relationship information of the capacitive current of the target node and the change relationship information of the capacitive current of other nodes;

[0301] determine a plurality of edge relationship weights corresponding to a plurality of detection points according to the edge relationship weight between the target node and other nodes and a preset algorithm; the preset algorithm includes: a distance algorithm and / or a symmetry algorithm.

[0302] Optionally, the processing unit 1002 is specifically configured to:

[0303] perform a first operation on each detection point in the plurality of detection points to obtain a plurality of first matrix operators; the first operation includes: determining a sum of values in the edge relationship weight of one detection point as the first matrix operator of the detection point;

[0304] determine a second matrix operator; a value in an Nth row and an Mth column in the second matrix operator is: a sum of values corresponding to the edge relationship of the Nth row and the Mth column in the detection point in addition to the N*M-1 detection points other than the Nth row and the Mth column in the plurality of detection points.

[0305] Optionally, the processing unit 1002 is specifically configured to:

[0306] According to the capacitance current change information of the screen and the integrated graph neural topology structure, a first value G and a second value K are determined; the first value G is used to represent a capacitance change matrix; and the second value K is used to represent a current change matrix.

[0307] The first value G and the second value K satisfy the following formula:

[0308] G = relue<sigmod[Q 综合 *C]>;

[0309] K = relue<sigmod[Q 综合 *I]>;

[0310] Wherein, Q 综合 is used to represent the integrated graph neural topology structure; C is used to represent the capacitance change information of the screen; I is used to represent the current change information of the screen; relue is used to represent an activation function; and sigmoid is used to represent a sigmoid function.

[0311] When the first value G is greater than a preset value, and / or the second value K is greater than a preset value, it is determined that the touch mode is the superimposed touch mode.

[0312] When the first value G is less than or equal to a preset value, and the second value K is less than or equal to a preset value, it is determined that the touch mode is the foreign matter touch mode or the non-foreign matter touch mode.

[0313] Optionally, the processing unit 1002 is further configured to, when the touch mode of the touch operation is the superimposed touch mode, determine a foreign matter content of the touch operation according to the capacitance current change information of the touch point corresponding to the touch operation, and determine a target foreign matter graph neural topology structure corresponding to the foreign matter content.

[0314] The processing unit 1002 is further configured to determine a first matrix X and a second matrix Y according to the non-foreign matter graph neural topology structure, the target foreign matter graph neural topology structure, the capacitance change matrix, the current change matrix, and an edge relationship weight of the touch point; the first matrix X and the second matrix Y satisfy the following formula:

[0315] X = FFT -1 [FFT(Q 非异物 )*FFT(G)] * sum(W);

[0316] Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)] * sum(W);

[0317] Wherein, FFT -1 is an inverse Fourier transform function, and FFT is a fast Fourier transform function; Q 目标异物W represents an edge relationship weight of the touch point;

[0318] The processing unit 1002 is further configured to not perform the capacitance compensation on the touch point when the first matrix X and the second matrix Y satisfy a target relationship.

[0319] (X2-X1) / X1>(Y2-Y1) / Y1.

[0320] X1 represents a value corresponding to the touch point in the first matrix X; X2 represents a value corresponding to a point other than the touch point in the first matrix X; Y1 represents a value corresponding to the touch point in the second matrix Y; and Y2 represents a value corresponding to a point other than the touch point in the second matrix Y.

[0321] The processing unit 1002 is further configured to adjust the reference capacitance of the touch point until all the touch points in the screen satisfy the target relationship when the first matrix X and the second matrix Y do not satisfy the target relationship.

[0322] The embodiment of the present application further provides a computer readable storage medium, which includes computer execution instructions. When the computer execution instructions run on a computer, the computer execution instructions make the computer execute the touch recognition method provided in the above embodiment.

[0323] The embodiment of the present application further provides a computer program, which can be directly loaded into a memory and includes software codes. The computer program can be loaded and executed by a computer to realize the touch recognition method provided in the above embodiment.

[0324] Those skilled in the art can realize that the functions described in the above one or more examples can be realized by hardware, software, firmware or any combination thereof. When realized by software, the functions can be stored in a computer readable medium or transmitted as one or more instructions or codes on a computer readable medium. The computer readable medium includes a computer readable storage medium and a communication medium, wherein the communication medium includes any medium that facilitates the transmission of computer programs from one place to another. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0325] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is taken as an example for illustration. In actual application, the above functions can be completed by different functional modules according to needs, i.e., the internal structure of the device is divided into different functional modules to complete all or part of the functions described above.

[0326] In several embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented by other manners. For example, the apparatus embodiments described above are merely illustrative, for example, the division of the modules or units is merely a logical function division, and actual implementation can have another division manner. For example, multiple units or components can be combined or integrated into another apparatus, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, apparatuses or units, which can be electrical, mechanical or other forms. The units described as separate components can be or can not be physically separated, and the components shown as units can be one physical unit or multiple physical units, that is, can be located in one place or can be distributed to multiple different places. Some or all units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0327] In addition, each functional unit in the embodiments of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions to make an apparatus (which can be a single chip, a chip, etc.) or a processor execute all or part of the steps of the method described in the embodiments of the present application. The storage medium mentioned above includes: a U disk, a mobile hard disk, a ROM, a RAM, a magnetic disk or an optical disk, and various program code storage media.

[0328] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A touch recognition method, characterized by, The method comprises the steps of: obtaining capacitance current change information of the screen in response to a touch operation performed by a user on the screen; determining a touch mode of the touch operation according to the capacitance current change information of the screen and a comprehensive graph neural topology structure constructed by a graph neural network; the comprehensive graph neural topology structure is used to represent a change relationship of capacitance currents of a plurality of detection points on the screen; the construction process of the comprehensive graph neural topology structure comprises the steps of: performing graph neural topology structure construction processing on a plurality of touch operations of the touch mode received by the screen to obtain a plurality of graph neural topology substructures corresponding to the plurality of touch operations one by one; the plurality of touch operations of the touch mode comprise touch operations of a non-foreign object touch mode and touch operations of a plurality of foreign object touch modes with different foreign object contents; the plurality of graph neural topology substructures comprise a non-foreign object graph neural topology structure and a plurality of foreign object graph neural topology structures corresponding to the plurality of different foreign object contents one by one; The graph neural topology structure construction process comprises: determining the capacitance current change information of each detection point in the plurality of detection points; determining the edge relationship weight between one detection point and other detection points according to the capacitance current change information of the one detection point, to obtain a plurality of edge relationship weights corresponding to the plurality of detection points one by one; determining a plurality of first matrix operators and a plurality of second matrix operators corresponding to the plurality of detection points one by one according to the plurality of edge relationship weights; and determining the graph neural topology substructure according to the plurality of first matrix operators and the plurality of second matrix operators. 子结构 satisfies the following formula: Q 子结构 = norm [sigmod {A 总 *B}] ; wherein sigmod is a preset function, Norm is a normalization function, A 总 is a sum of the plurality of first matrix operators, and B is the second matrix operator. a first matrix operator of one detection point is used to represent an adjacent relationship between the one detection point and other detection points; and a second matrix operator is used to represent superimposed effects of the other detection points on the one detection point; the comprehensive graph neural topology structure is determined according to the plurality of graph neural topology substructures and convolution operation; the comprehensive graph neural topology structure satisfies the following formula: Q 综合 = Q 非异物 * Q 异物 ; wherein Q 综合 for representing the integrated graph neural topology; Q 非异物 for representing the non-foreign matter graph neural topology; Q 异物 for representing one foreign matter graph neural topology; for representing the convolution operation.

2. The touch recognition method according to claim 1, wherein, the touch mode comprises any one of a foreign object touch mode, a non-foreign object touch mode, and a superimposed touch mode of the foreign object touch mode and the non-foreign object touch mode.

3. The touch recognition method according to claim 2, wherein, The determination of the capacitance current change information of each detection point in the plurality of detection points comprises the steps of: dividing the screen into an N×M horizontal and vertical array to obtain N×M nodes; N and M are positive integers; one node corresponds to one detection point; determining a target node in the N×M nodes; the target node comprises a center point of the N×M nodes, an upper vertex of a central axis of the N×M nodes, a left vertex of the N×M nodes, and a left vertex of the central axis of the N×M nodes; performing a touch operation on the target node to obtain change relationship information of a capacitance current of the target node and change relationship information of capacitance currents of other nodes.

4. The touch recognition method according to claim 3, wherein The determination of the edge relationship weight between the one detection point and the other detection points according to the capacitance current change information of the one detection point comprises the steps of: determining the edge relationship weight between the target node and the other nodes according to the change relationship information of the capacitance current of the target node and the change relationship information of the capacitance currents of the other nodes; determining a plurality of edge relationship weights corresponding to the plurality of detection points according to the edge relationship weight between the target node and the other nodes and a preset algorithm; the preset algorithm comprises a distance algorithm and / or a symmetry algorithm.

5. The touch recognition method according to claim 4, wherein, The determination of the plurality of first matrix operators and the second matrix operator corresponding to the plurality of detection points according to the plurality of edge relationship weights comprises the steps of: performing a first operation on each of the plurality of detection points to obtain the plurality of first matrix operators; the first operation comprises: determining the sum of the values in the edge relationship weight of the one detection point as the first matrix operator of the one detection point; determining the second matrix operator; the value of the Nth row and the Mth column in the second matrix operator is: the sum of the values in the edge relationship weight of the N×M-1 detection points corresponding to the detection point in the Nth row and the Mth column in the plurality of detection points, except for the detection point in the Nth row and the Mth column in the plurality of detection points.

6. The touch recognition method according to claim 2, wherein, The determination of the touch mode of the touch operation according to the capacitive current change information of the screen and the pre-constructed integrated graph neural topology structure comprises: determining a first value G and a second value K according to the capacitive current change information of the screen and the integrated graph neural topology structure; the first value G is used to represent a capacitive change matrix; the second value K is used to represent a current change matrix; The first value G and the second value K satisfy the following formula: G = relue < sigmod[Q 综合 C] ; K = relue < sigmod[Q 综合 I] ; wherein Q 综合 for representing the integrated graph neural topology; C for representing the capacitance change information of the screen; I for representing the current change information of the screen; relue for representing an activation function; sigmoid for representing a sigmoid function; When the first value G is greater than a preset value, and / or the second value K is greater than the preset value, it is determined that the touch mode is the superimposed touch mode; When the first value G is less than or equal to the preset value, and the second value K is less than or equal to the preset value, it is determined that the touch mode is the foreign matter touch mode or the non-foreign matter touch mode.

7. The touch recognition method according to claim 6, wherein, Further comprising: When the touch mode of the touch operation is the superimposed touch mode, determining the foreign matter content of the touch operation according to the capacitive current change information of the touch point corresponding to the touch operation, and determining a target foreign matter graph neural topology structure corresponding to the foreign matter content; determining a first matrix X and a second matrix Y according to the non-foreign matter graph neural topology structure, the target foreign matter graph neural topology structure, the capacitive change matrix, the current change matrix and the edge relationship weight of the touch point; the first matrix X and the second matrix Y satisfy the following formula: X = FFT -1 [FFT(Q 非异物 )*FFT(G)]*sum(W); Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)]*sum(W); wherein FFT is a Fourier transform function, and FFT is a fast Fourier transform function. -1 wherein FFT is a Fourier transform function, and FFT is a fast Fourier transform function. 目标异物 for representing the target foreign matter graph neural topology structure; and W for representing the edge relationship weight of the touch point. When the first matrix X and the second matrix Y satisfy a target relationship, no capacitive compensation is performed on the touch point; the target relationship is: (X2-X1) / X1>(Y2-Y1) / Y1; Wherein, X1 is used to represent the value corresponding to the touch point in the first matrix X; X2 is used to represent the value corresponding to other points except the touch point in the first matrix X; Y1 is used to represent the value corresponding to the touch point in the second matrix Y; Y2 is used to represent the value corresponding to other points except the touch point in the second matrix Y; When the first matrix X and the second matrix Y do not satisfy the target relationship, the reference capacitance of the touch point is adjusted until the touch points in the screen all satisfy the target relationship.

8. A touch recognition device, characterized by, Comprise: an acquisition unit and a processing unit; The acquisition unit is configured to acquire capacitive current change information of a screen in response to a touch operation performed by a user on the screen. The processing unit is configured to determine a touch mode of the touch operation according to the change information of the capacitive current of the screen and a comprehensive graph neural topology constructed by a graph neural network. The comprehensive graph neural topology is constructed in the following procedure: The graph neural topology structure construction processing is performed on the touch operation of each of the plurality of touch modes received by the screen, to obtain a plurality of graph neural topology substructures corresponding to the plurality of touch operations one by one. The plurality of touch modes include a non-foreign object touch mode and a plurality of foreign object touch modes with different foreign object contents. The graph neural topology structure construction process comprises: determining the capacitance current change information of each detection point in the plurality of detection points; determining the edge relationship weight between one detection point and other detection points according to the capacitance current change information of the one detection point, to obtain a plurality of edge relationship weights corresponding to the plurality of detection points one by one; determining a plurality of first matrix operators and a plurality of second matrix operators corresponding to the plurality of detection points one by one according to the plurality of edge relationship weights; and determining the graph neural topology substructure according to the plurality of first matrix operators and the plurality of second matrix operators. 子结构 satisfies the following formula: Q 子结构 = norm [sigmod {A 总 *B}] ; wherein sigmod is a preset function, Norm is a normalization function, A 总 is a sum of the plurality of first matrix operators, and B is the second matrix operator. The plurality of graph neural topology substructures include a non-foreign object graph neural topology structure and a plurality of foreign object graph neural topology structures corresponding to the plurality of different foreign object contents one by one. The first matrix operator of the one detection point is used to represent the adjacent relationship between the one detection point and the other detection points. Q 综合 = Q 非异物 * Q 异物 ; wherein Q 综合 for representing the integrated graph neural topology; Q 非异物 for representing the non-foreign graph neural topology; Q 异物 for representing one foreign graph neural topology; *for representing the convolution operation.

9. The touch recognition apparatus of claim 8, wherein, The comprehensive graph neural topology is determined according to the plurality of graph neural topology substructures and a convolution operation.

10. The touch recognition apparatus of claim 9, wherein, The touch mode includes any one of a foreign object touch mode, a non-foreign object touch mode, and a superimposed touch mode of the foreign object touch mode and the non-foreign object touch mode. The processing unit is specifically configured to: divide the screen into an N×M horizontal and vertical array to obtain N×M nodes; N and M are positive integers; one node corresponds to one detection point; determine a target node in the N×M nodes; the target node includes a center point of the N×M nodes, an upper vertex of a central axis of the N×M nodes, a left vertex of the N×M nodes, and a left vertex of the central axis of the N×M nodes; 11. The touch recognition apparatus of claim 10, wherein, perform a touch operation on the target node to obtain change relationship information of the capacitive current of the target node and change relationship information of the capacitive current of other nodes. The processing unit is specifically configured to: determine an edge relationship weight between the target node and the other nodes according to the change relationship information of the capacitive current of the target node and the change relationship information of the capacitive current of the other nodes; determine a plurality of edge relationship weights corresponding to the plurality of detection points according to the edge relationship weight between the target node and the other nodes and a preset algorithm.

12. The touch recognition apparatus of claim 11, wherein, The preset algorithm includes a distance algorithm and / or a symmetry algorithm. The processing unit is specifically configured to: perform a first operation on each of the plurality of detection points to obtain the plurality of first matrix operators; the first operation includes determining a sum of values in the edge relationship weight of the one detection point as the first matrix operator of the one detection point. A value in an Nth row and an Mth column of the second matrix operator is a sum of values of edge relationship weights of N×M-1 detection points other than an Nth row and an Mth column of the plurality of detection points, which have an edge relationship with the Nth row and the Mth column of the detection points.

13. The touch recognition apparatus of claim 9, wherein, The processing unit is specifically configured to: determine a first value G and a second value K according to the capacitive current change information of the screen and the integrated graph neural topology structure; the first value G is used to represent a capacitance change matrix; and the second value K is used to represent a current change matrix; the first value G and the second value K satisfy the following formula: G = relue < sigmod[Q 综合 C] ; K = relue < sigmod[Q 综合 I] ; wherein Q 综合 for representing the integrated graph neural topology; C for representing the capacitance change information of the screen; I for representing the current change information of the screen; relue for representing an activation function; sigmoid for representing a sigmoid function; when the first value G is greater than a preset value, and / or the second value K is greater than the preset value, it is determined that the touch mode is the superimposed touch mode; when the first value G is less than or equal to the preset value, and the second value K is less than or equal to the preset value, it is determined that the touch mode is the foreign matter touch mode or the non-foreign matter touch mode.

14. The touch recognition apparatus according to claim 13, wherein the processing unit is further configured to, when the touch mode of the touch operation is the superimposed touch mode, determine a foreign matter content of the touch operation according to capacitive current change information of a touch point corresponding to the touch operation, and determine a target foreign matter graph neural topology structure corresponding to the foreign matter content; the processing unit is further configured to determine a first matrix X and a second matrix Y according to the non-foreign matter graph neural topology structure, the target foreign matter graph neural topology structure, the capacitance change matrix, the current change matrix, and an edge relationship weight of the touch point; the first matrix X and the second matrix Y satisfy the following formula: X = FFT -1 [FFT(Q 非异物 )*FFT(G)]*sum(W); Y = FFT -1 [FFT(Q 目标异物 )*FFT(G)]*sum(W); wherein FFT is a Fourier transform function, and FFT is a fast Fourier transform function. -1 wherein FFT is a Fourier transform function, and FFT is a fast Fourier transform function. 目标异物 for representing the target foreign matter graph neural topology structure; and W for representing the edge relationship weight of the touch point. the processing unit is further configured to, when the first matrix X and the second matrix Y satisfy a target relationship, not perform capacitive compensation on the touch point; the target relationship is: (X2-X1) / X1>(Y2-Y1) / Y1; wherein X1 represents a value corresponding to the touch point in the first matrix X; X2 represents a value corresponding to a point other than the touch point in the first matrix X; Y1 represents a value corresponding to the touch point in the second matrix Y; and Y2 represents a value corresponding to a point other than the touch point in the second matrix Y; the processing unit is further configured to, when the first matrix X and the second matrix Y do not satisfy the target relationship, adjust a reference capacitance of the touch point until all touch points in the screen satisfy the target relationship.

15. A touch recognition device, characterized by A touch recognition apparatus includes a memory and a processor; the memory is used to store computer execution instructions; the processor is connected with the memory through a bus; when the touch recognition apparatus is running, the processor executes the computer execution instructions stored in the memory, so that the touch recognition apparatus performs a touch recognition method according to any one of claims 1-7.

16. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises computer-executable instructions that, when executed on a computer, cause the computer to perform the touch recognition method according to any one of claims 1-7.

Images