CN-121979405-A - Touch object type identification method and device, terminal equipment and storage medium

CN121979405ACN 121979405 ACN121979405 ACN 121979405ACN-121979405-A

Abstract

The application is applicable to the technical field of intelligent terminals and provides a method, a device, terminal equipment and a storage medium for identifying the type of a touch object, wherein the method comprises the steps of responding to touch operation in an infrared touch area to obtain position information of a touch point; the method comprises the steps of obtaining shielding information corresponding to touch operation in an infrared touch area, constructing a target feature vector according to the position information and the shielding information, inputting the target feature vector into a pre-trained object type recognition model to obtain an object type and corresponding confidence coefficient output by the object type recognition model, and determining a touch object type corresponding to a touch point according to the object type and the corresponding confidence coefficient. The method and the device can improve the accuracy and stability of touch object type identification, and ensure the high efficiency of the identification process to adapt to real-time touch requirements.

Inventors

LIU JIGUI
WANG SHUAI
ZHANG MINGHUA
Chen Rijie
CHEN FUBIN

Assignees

深圳市鸿合创新信息技术有限责任公司

Dates

Publication Date: 20260505
Application Date: 20251224

Claims (10)

1. The touch object type identification method is characterized by comprising the following steps of: Responding to touch operation in the infrared touch area, and obtaining position information of a touch point; acquiring shielding information corresponding to the touch operation in the infrared touch area; constructing a target feature vector according to the position information and the shielding information; inputting the target feature vector into a pre-trained object type recognition model to obtain the object type and the corresponding confidence coefficient output by the object type recognition model; and determining the touch object type corresponding to the touch point according to the object type and the corresponding confidence level.
2. The method according to claim 1, wherein the step of obtaining occlusion information corresponding to the touch operation in the infrared touch area includes: determining initial shielding information according to the infrared beam information shielded by the touch operation; when the touch point is positioned at the edge or corner area of the infrared touch area, determining weighting coefficients corresponding to infrared light beams in different incidence directions according to the position information; carrying out fusion processing on the information of the blocked infrared light beams from different incidence directions by utilizing the weighting coefficient; And carrying out weighted correction on the initial shielding information according to the fusion processing result to obtain corrected shielding information.
3. The method of claim 1, wherein determining the touch object type corresponding to the touch point according to the object type and the confidence level thereof comprises: If the confidence coefficient is higher than or equal to a preset confidence coefficient threshold value, determining the object type as a touch object type corresponding to the touch operation; And if the confidence coefficient is lower than the preset confidence coefficient threshold value, determining a default object type as the touch object type corresponding to the touch operation.
4. The method of claim 1, wherein determining the touch object type corresponding to the touch point according to the object type and the confidence level thereof comprises: If the confidence coefficient is lower than a preset confidence coefficient threshold value, correcting the shielding information according to the position information of the touch point; reconstructing a target feature vector according to the position information and the corrected shielding information; inputting the reconstructed target feature vector into the object type recognition model to obtain the object type and the corresponding confidence coefficient of the object type newly output by the object type recognition model; and determining the touch object type corresponding to the touch point according to the newly output object type and the corresponding confidence coefficient.
5. The method of claim 1, wherein determining the touch object type corresponding to the touch point according to the object type and the confidence level thereof comprises: acquiring a continuous touch sequence, wherein the continuous touch sequence comprises the touch points and a plurality of historical touch points adjacent to the touch points in historical time sequence; acquiring a history identification result corresponding to a history touch point in the continuous touch sequence; And carrying out voting decision according to the historical recognition result, and determining the touch object type corresponding to the touch point based on the voting decision result.
6. The method according to any one of claims 1 to 5, wherein the pre-trained object type recognition model is obtained by: Collecting touch samples of a plurality of touch objects at different positions of an infrared touch area, wherein each touch sample comprises sample position information, sample shielding information and corresponding object type labels; Performing position normalization processing on the sample shielding information according to the sample position information to obtain corrected sample shielding information; constructing a sample feature vector according to the sample position information and the corrected sample shielding information; And carrying out iterative training on the initial object type recognition model by using the sample feature vector and the corresponding object type label until the model converges, and obtaining the object type recognition model which is trained in advance.
7. The method of claim 6, wherein the step of iteratively training an initial object type recognition model using the sample feature vectors and corresponding object type tags further comprises: determining a weighted loss function, wherein the weight value of a touch sample positioned at the edge and corner area of the infrared touch area is higher than that of a touch sample positioned at the central area of the infrared touch area; and optimizing the initial object type identification model by adopting the weighted loss function.
8. A touch object type recognition device, characterized by comprising: the first information acquisition unit is used for responding to touch operation in the infrared touch area and acquiring position information of a touch point; the second information acquisition unit is used for acquiring shielding information corresponding to the touch operation in the infrared touch area; the feature vector construction unit is used for constructing a target feature vector according to the position information and the shielding information; the type information identification unit is used for inputting the target feature vector into a pre-trained object type identification model to obtain the object type and the corresponding confidence coefficient output by the object type identification model; and the object type determining unit is used for determining the touch object type corresponding to the touch point according to the object type and the corresponding confidence level thereof.
9. Terminal device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method for recognizing a touch object type according to any of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the touch object type recognition method according to any one of claims 1 to 7.

Description

Touch object type identification method and device, terminal equipment and storage medium Technical Field The present application relates to the field of intelligent terminals, and in particular, to a method and apparatus for identifying a touch object type, a terminal device, and a storage medium. Background The infrared touch technology has been widely used in electronic whiteboard, interactive display screen, conference projection and other devices due to the advantages of controllable cost, wide application scene and the like, and the touch point positioning is realized by shielding infrared light paths around the screen by a touch object. With the richness of application scenes of the touch control device, users have put higher demands on the intelligence of the touch control device, and hope that the touch control device can automatically identify the types of objects used in the current touch control operation, such as distinguishing fingers, touch control pens, blackboard erasers and the like, so as to trigger corresponding intelligent interaction functions. Currently, a common touch object type recognition scheme mainly depends on simple measurement and threshold judgment of an infrared shielding signal. However, because the infrared light paths are not uniformly distributed in the plane of the screen, the recognition accuracy of the judgment rules based on the fixed threshold or the single measurement characteristic in different areas of the screen is low, and the stability of the overall recognition is insufficient. And by adopting a scheme of complex image processing, shape fitting or multiple re-projection calculation, the system response delay or high calculation resource consumption can be caused, and the real-time touch control requirement is difficult to adapt. In view of this, how to improve accuracy and stability of touch object type recognition, and ensure high efficiency of recognition process to adapt to real-time touch requirements at the same time is a technical problem to be solved at present. Disclosure of Invention The embodiment of the application provides a touch object type identification method, a device, terminal equipment and a storage medium, which can improve the accuracy and stability of touch object type identification, and ensure the high efficiency of an identification process to adapt to real-time touch requirements. In a first aspect, an embodiment of the present application provides a method for identifying a touch object type, including: Responding to touch operation in the infrared touch area, and obtaining position information of a touch point; acquiring shielding information corresponding to the touch operation in the infrared touch area; constructing a target feature vector according to the position information and the shielding information; inputting the target feature vector into a pre-trained object type recognition model to obtain the object type and the corresponding confidence coefficient output by the object type recognition model; and determining the touch object type corresponding to the touch point according to the object type and the corresponding confidence level. In a possible implementation manner of the first aspect, the step of obtaining occlusion information corresponding to the touch operation in the infrared touch area includes: determining initial shielding information according to the infrared beam information shielded by the touch operation; when the touch point is positioned at the edge or corner area of the infrared touch area, determining weighting coefficients corresponding to infrared light beams in different incidence directions according to the position information; carrying out fusion processing on the information of the blocked infrared light beams from different incidence directions by utilizing the weighting coefficient; And carrying out weighted correction on the initial shielding information according to the fusion processing result to obtain corrected shielding information. In a possible implementation manner of the first aspect, the determining, according to the object type and the confidence level corresponding to the object type, a touch object type corresponding to the touch point includes: If the confidence coefficient is higher than or equal to a preset confidence coefficient threshold value, determining the object type as a touch object type corresponding to the touch operation; And if the confidence coefficient is lower than the preset confidence coefficient threshold value, determining a default object type as the touch object type corresponding to the touch operation. In a possible implementation manner of the first aspect, the determining, according to the object type and the confidence level corresponding to the object type, a touch object type corresponding to the touch point includes: If the confidence coefficient is lower than a preset confidence coefficient threshold value, correcting the shielding information