Method, device, and system for determining a false touch on a touch screen of an electronic device using an AI model

A computer readable recording medium storing a program programmed to determine a false touch on a touch screen of an electronic apparatus, the program having computer-executable instructions for performing steps comprising:

• detecting a touch input in order to obtain a touch result value through the touch screen;
• detecting at least two of a time interval with a previously input touch input of the detected touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input;
• inputting, to a computer configured to execute a learned artificial intelligence (AI) model, data related to the detected at least two of the time interval, the touch area, the touch intensity, or the temperature to determine whether a touch type of the detected touch input is a normal touch or a false touch;
• applying the input data to the learned AI model;
• and
• outputting, by the learned AI model, whether the detected touch input is a normal touch or a false touch,
• wherein the normal touch is a touch that a target setting value of an intended touch input and the touch result value of the detected touch input coincide with each other, and the false touch is a touch that an unintended touch result value is detected even without touch input, or the target setting value and the touch result value of the detected touch input does not coincide with each other,
• wherein the learned AI model is a reinforcement learning based neural network model configured so that a false touch determining program for determining whether a touch type of the touch input is a normal touch or a false touch performs an action for deciding the touch type with respect to a state where at least two of a time interval of the touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input have been given, and the program receives a reward based on the decided touch type being an intended touch type, and
• wherein the reinforcement learning based neural network model is configured so that the false touch determining program performs an action in the state and updates a false touch determining performance in order to receive a maximum reward, in a false touch determining environment that receives the state and the action of the program to return the reward and a next state.

A false touch determining device for determining a false touch on a touch screen of an electronic apparatus, comprising:

• a circuit configured to detect a touch input to obtain a touch result value through the touch screen, and detect at least two of a time interval of the detected touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input;
• a computer configured to execute a learned artificial intelligence (AI) model;
• and
• an output circuit configured to output whether the detected touch input is a normal touch or a false touch from the learned AI model by applying, to the learned AI model, data related to at least two of the time interval of the detected touch input, the touch area of the touch input, the touch intensity of the touch input, or the temperature of the touch input to determine whether a touch type of the detected touch input is a normal touch or a false touch,
• wherein the learned AI model is a reinforcement learning based neural network model configured so that a program for determining whether a touch type of the touch input is a normal touch or a false touch performs an action for deciding the touch type with respect to a state where at least two of a time interval of the touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input have been given, and the program receives a reward based on the decided touch type being an intended touch type, and
• wherein the reinforcement learning based neural network model is configured so that the program performs an action in the state and updates a false touch determining performance in order to receive a maximum reward, in an environment that obtains the state and the action of the program to return the reward and a next state.

A touch screen system comprising a false touch determining device for determining a false touch on a touch screen of an electronic apparatus and a server, the touch screen system comprising:

• a circuit configured to detect a touch input to obtain a touch result value through the touch screen, and detect at least two of a time interval of the detected touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input;
• a computer configured to execute a learned artificial intelligence (AI) model;
• an output circuit configured to output whether the detected touch input is a normal touch or a false touch from the learned AI model by applying, to the learned AI model, data related to at least two of the time interval of the detected touch input, the touch area of the touch input, the touch intensity of the touch input, or the temperature of the touch input to determine whether a touch type of the detected touch input is a normal touch or a false touch;
• and
• a transceiver configured to communicate with the server, the transceiver configured to transmit, to the server, the detected at least two of the time interval of a plurality of the touch inputs, the touch area of the touch input, the touch intensity of the touch input, or the temperature of the touch input, and touch types labeled to the touch input,
• wherein the server is configured to execute the AI model learner for generating the learned AI model that has learned at least two of the plurality of the touch inputs, the touch area of the touch input, the touch intensity of the touch input, or the temperature of the touch input, which have been received, and the touch types labeled to the touch input through a deep neural network,
• wherein the server is configured to transmit, to the false touch determining device, the learned AI model that has learned through the AI model learner, and
• wherein the output circuit of the false touch determining device is configured to infer whether a touch type of the detected touch input is a normal touch or a false touch through the learned AI model received from the server,
• wherein the learned AI model is a reinforcement learning based neural network model configured so that a program for determining whether a touch type of the touch input is a normal touch or a false touch performs an action for deciding the touch type with respect to a state where at least two of a time interval of the touch input, a touch area of the touch input, a touch intensity of the touch input, or a temperature of the touch input have been given, and the program receives a reward based on the decided touch type being an intended touch type, and
• wherein the reinforcement learning based neural network model is configured so that the program performs an action in the state and updates a false touch determining performance in order to receive a maximum reward, in an environment that obtains the state and the action of the program to return the reward and a next state.