Zebra crossing passing scene shooting picture recognition system
By obtaining multiple scene parameters and basic data of image processing types of zebra crossing scenes, using customized AI recognition models and continuous learning components to train echo state networks, the problem of difficult image authenticity in zebra crossing monitoring is solved, and the passage of zebra crossings is achieved Accurate monitoring of scenes and high-quality image recognition.
Patent Information
- Application Number
- CN202411053422.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The prior art is difficult to ensure the authenticity of the monitoring images in zebra crossing monitoring, especially the quality reduction caused by image processing and transmission, which affects the accurate monitoring of zebra crossing scenes.
By obtaining multiple scene parameters and basic data of the zebra crossing scene, a customized AI recognition model is used for intelligent analysis, and the echo state network is trained in combination with continuous learning components to judge the authenticity of the image and identify the image processing type.
It realizes intelligent judgment of the authenticity of the shooting pictures of the zebra crossing scene, provides reliable reference data, and ensures the quality and accuracy of the monitoring image.
Smart Images

Figure CN118942055A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of zebra crossing monitoring, and in particular to a zebra crossing traffic scene shooting picture recognition system. Background Art
[0002] There is an international standard for the width, length and interval of zebra crossings. The basic length of a zebra crossing is 3 to 5 meters, and it should cross the road outside the sidewalk. The width of each zebra crossing is 45 or 60 centimeters. The width of the zebra crossing can also be widened by 1 meter according to the number of pedestrians. The more pedestrians there are, the wider the zebra crossing will be. On a section of road, the interval between zebra crossings is generally 150 to 500 meters, and the more reasonable interval is 350 to 400 meters. That is to say, in the city, it is best to set up a zebra crossing every 350 to 400 meters to facilitate pedestrian passage. If there are no pedestrian crossing facilities on the roads in front of schools, kindergartens, hospitals and nursing homes, pedestrian crossings should be marked and warning signs should be set up. The sidewalks of major urban roads should be equipped with blind paths according to the plan.
[0003] As an important scene of traffic monitoring, the monitoring and management of zebra crossings has very high requirements on the authenticity of the monitoring images. However, in the actual image acquisition process, due to compression requirements and / or transmission requirements, the original image after shooting will be processed by multiple image processing types before the monitoring image obtained by the monitoring end can be obtained. Therefore, it is necessary to perform content analysis on the monitoring image to identify the multiple image processing types it has passed through, which provides an important basis for the true reproduction of the zebra crossing traffic scene. Summary of the invention
[0004] In order to overcome the technical problems in the prior art, the present invention proposes a zebra crossing scene shooting picture recognition system, which obtains the imaging depths of field corresponding to each edge pixel point of the main object in the current shooting picture and the coordinate values corresponding to each edge pixel point, so as to obtain the basic data required for the intelligent analysis of multiple image processing types executed on the current shooting picture, wherein the main object in the current shooting picture is the object occupying the largest area in the current shooting picture, and multiple scene parameters of the scene where the zebra crossing is located, the imaging depths of field corresponding to each edge pixel point of the main object in the current shooting picture, the coordinate values and the scene shooting components are obtained. Multiple working parameters of the photoelectric sensing component are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the type names represented by multiple binary numerical values corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the model, thereby completing the intelligent judgment of the authenticity of the acquired current shooting picture, and using a customized AI recognition model for intelligent analysis of the multiple image processing types executed on the current shooting picture. The structural customization of the AI recognition model lies in the introduction of a continuous learning component for training the echo state network for each training to obtain the echo state network after each training and output it as the AI recognition model.
[0005] According to the present invention, a zebra crossing scene shooting picture recognition system is provided, the system comprising: The scene shooting component is located above the zebra crossing and is used to shoot the scene where the zebra crossing is located when the signal light corresponding to the zebra crossing is in the green light on time period, so as to obtain and output the corresponding current shooting picture; A continuous learning component, used to train the echo state network for each time to obtain the echo state network after each training and output it as an AI recognition model; A touch display screen connected to the customized identification mechanism, used to receive and display a plurality of binary numerical type names corresponding to a plurality of image processing types that have been executed in the scene capture component for the current captured image; A parameter parsing mechanism, used for parsing a plurality of scene parameters of the scene where the zebra crossing is located, wherein the plurality of scene parameters of the scene where the zebra crossing is located are a zebra crossing distribution length, a zebra crossing distribution area, and an area occupied by a traffic intersection where the zebra crossing is located in the scene where the zebra crossing is located; A customized recognition mechanism is connected to the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively, and is used to synchronously input multiple scene parameters of the scene where the zebra crossing is located, various imaging depths of field corresponding to various edge pixel points of the main object in the current shooting picture, various coordinate values, and multiple working parameters of the photoelectric sensor component of the scene shooting component into the AI recognition model, and execute the AI recognition model to obtain the type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model; Among them, each imaging depth of field corresponding to each edge pixel point of the main object in the current shooting picture, each coordinate value, multiple working parameters of the photoelectric sensor component of the scene shooting component, and the number of noise types of the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The multiple binary numerical values corresponding to the distribution of multiple image processing types executed in the scene shooting component include: the multiple working parameters of the photoelectric sensor component of the scene shooting component are the electronic shutter rate, frame rate, signal-to-noise ratio and quantum efficiency of the photoelectric sensor component of the scene shooting component; Among them, the imaging depths of field, the coordinate values, the multiple working parameters of the photoelectric sensor components of the scene shooting components and the number of noise types in the current shooting picture corresponding to the edge pixel points of the main object in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component also include: the main object in the current shooting picture is the object occupying the largest area in the current shooting picture.
[0006] It can be seen that the present invention has at least the following three important inventive features: Important invention point A: obtaining each imaging depth of field corresponding to each edge pixel point of the main object in the current shooting picture and each coordinate value corresponding to each edge pixel point, so as to obtain basic data required for intelligent analysis of multiple image processing types executed on the current shooting picture, wherein the main object in the current shooting picture is the object occupying the largest area in the current shooting picture; Important invention point B: Multiple scene parameters of the scene where the zebra crossing is located, various imaging depths of field corresponding to each edge pixel point of the main object in the current shooting picture, various coordinate values, and multiple working parameters of the photoelectric sensor component of the scene shooting component are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model, thereby completing the intelligent judgment of the authenticity of the acquired current shooting picture; Important Invention Point C: A customized AI recognition model is used to intelligently analyze multiple image processing types that have been performed on the current shooting picture. The structural customization of the AI recognition model lies in the introduction of a continuous learning component for training the echo state network multiple times to obtain the echo state network after each training and output it as the AI recognition model.
[0007] The zebra crossing scene shooting picture recognition system of the present invention is intelligently designed and widely used. Since it can use an AI recognition model to intelligently recognize multiple types of image processing performed on the current shooting picture based on multiple visual information of the current shooting picture, it can provide reliable reference data for the true restoration of the zebra crossing scene. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein: Figure 1 1 is a block diagram of the internal structure of a zebra crossing scene shooting picture recognition system according to Embodiment 1 of the present invention.
[0009] Figure 2 1 is a block diagram of the internal structure of a zebra crossing scene shooting picture recognition system according to Embodiment 2 of the present invention.
[0010] Figure 3 1 is a block diagram of the internal structure of a zebra crossing scene shooting picture recognition system according to Embodiment 3 of the present invention. DETAILED DESCRIPTION
[0011] The implementation scheme of the zebra crossing scene shooting picture recognition system of the present invention will be described in detail with reference to the accompanying drawings.
[0012] Figure 1 The internal structure block diagram of the zebra crossing scene shooting picture recognition system according to Embodiment 1 of the present invention is shown, and the system includes: The scene shooting component is located above the zebra crossing and is used to shoot the scene where the zebra crossing is located when the signal light corresponding to the zebra crossing is in the green light on time period, so as to obtain and output the corresponding current shooting picture; A continuous learning component, used to train the echo state network for each time to obtain the echo state network after each training and output it as an AI recognition model; Specifically, the continuous learning component is used to train the echo state network for each time to obtain the echo state network after each training and output it as the AI recognition model, including: the number of times the echo state network is trained is monotonically positively correlated with the resolution of the scene capture component; A touch display screen connected to the customized identification mechanism, used to receive and display a plurality of binary numerical type names corresponding to a plurality of image processing types that have been executed in the scene capture component for the current captured image; A parameter parsing mechanism, used for parsing a plurality of scene parameters of the scene where the zebra crossing is located, wherein the plurality of scene parameters of the scene where the zebra crossing is located are a zebra crossing distribution length, a zebra crossing distribution area, and an area occupied by a traffic intersection where the zebra crossing is located in the scene where the zebra crossing is located; A customized recognition mechanism is connected to the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively, and is used to synchronously input multiple scene parameters of the scene where the zebra crossing is located, various imaging depths of field corresponding to various edge pixel points of the main object in the current shooting picture, various coordinate values, and multiple working parameters of the photoelectric sensor component of the scene shooting component into the AI recognition model, and execute the AI recognition model to obtain the type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model; Among them, each imaging depth of field corresponding to each edge pixel point of the main object in the current shooting picture, each coordinate value, multiple working parameters of the photoelectric sensor component of the scene shooting component, and the number of noise types of the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The multiple binary numerical values corresponding to the distribution of multiple image processing types executed in the scene shooting component include: the multiple working parameters of the photoelectric sensor component of the scene shooting component are the electronic shutter rate, frame rate, signal-to-noise ratio and quantum efficiency of the photoelectric sensor component of the scene shooting component; Among them, each imaging depth of field corresponding to each edge pixel point of the main object in the current shooting picture, each coordinate value, multiple working parameters of the photoelectric sensor component of the scene shooting component, and the number of noise types in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The multiple binary numerical values corresponding to the distribution of multiple image processing types executed in the scene shooting component also include: the main object in the current shooting picture is the object with the largest area in the current shooting picture; And wherein, the imaging depths of field, the coordinate values, the multiple working parameters of the photoelectric sensor component of the scene shooting component and the number of noise types of the current shooting picture corresponding to the edge pixel points of the main object in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the type names of the multiple binary numerical representations corresponding to the distribution of the multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model, which also includes: the type names of the multiple binary numerical representations corresponding to the distribution of the multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model are the representation mode of multiple binary numerical values connected end to end.
[0013] Figure 2 1 is a block diagram of the internal structure of a zebra crossing scene shooting picture recognition system according to Embodiment 2 of the present invention.
[0014] Compared to Figure 1 , Figure 2 The zebra crossing scene shooting picture recognition system in the present invention may also include the following components: a cache execution device, which is arranged near the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is connected to the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively; Among them, the cache execution device is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the cache execution device is used to respectively implement the cache operation of the input signals and output signals of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component.
[0015] Figure 3 1 is a block diagram of the internal structure of a zebra crossing scene shooting picture recognition system according to Embodiment 3 of the present invention.
[0016] Compared to Figure 1 , Figure 3 The zebra crossing scene shooting picture recognition system in the present invention may also include the following components: A positioning server component is arranged near the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is connected to the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively; Among them, the positioning server component is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the positioning server component is used to provide positioning data services based on the Beidou navigation mechanism for the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively.
[0017] Next, the specific structure of the zebra crossing scene shooting picture recognition system of the present invention will be further described.
[0018] In the zebra crossing scene shooting picture recognition system according to any embodiment of the present invention: A timing server component is used which is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, and is used to provide the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component with their respective required timing services.
[0019] In the zebra crossing scene shooting picture recognition system according to any embodiment of the present invention: A timing server device is provided near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component for providing the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component with their respective required timing services, including: the timing server device is an FPGA device.
[0020] In the zebra crossing scene shooting picture recognition system according to any embodiment of the present invention: A serial port configuration device is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, for providing on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component.
[0021] In the zebra crossing scene shooting picture recognition system according to any embodiment of the present invention: A serial port configuration device is provided near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component for providing on-site configuration operations of working parameters based on serial port connection for each of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the serial port configuration device is a programmable logic device.
[0022] And in the zebra crossing scene shooting picture recognition system according to any embodiment of the present invention: A serial port configuration device is used which is arranged near the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component to provide on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component. The serial port configuration device also includes: the serial port configuration device uses different configuration addresses to provide on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component.
[0023] In addition, in the zebra crossing scene shooting picture recognition system, the imaging depths of field, coordinate values, multiple working parameters of the photoelectric sensor component of the scene shooting component and the number of noise types in the current shooting picture corresponding to each edge pixel point of the main object in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The type names of multiple binary numerical representations corresponding to the distribution of multiple image processing types executed in the scene shooting component also include: using various model data of the AI recognition model to complete the model storage of the AI recognition model.
[0024] The present invention can be implemented and practiced in different forms without departing from its spirit and main features. Therefore, the embodiments described are considered to be exemplary in all respects without limiting the meaning. The scope of the present invention is limited by the appended claims, rather than by the above description. It is intended to include all modifications and variations within the equivalent scope of the appended claims.
Claims
1. A zebra crossing scene shooting picture recognition system, characterized in that: The system comprises: The scene shooting component is located above the zebra crossing and is used to shoot the scene where the zebra crossing is located when the signal light corresponding to the zebra crossing is in the green light on time period, so as to obtain and output the corresponding current shooting picture; A continuous learning component, used to train the echo state network for each time to obtain the echo state network after each training and output it as an AI recognition model; A touch display screen connected to the customized identification mechanism, used to receive and display a plurality of binary numerical type names corresponding to a plurality of image processing types that have been executed in the scene capture component for the current captured image; A parameter parsing mechanism, used for parsing a plurality of scene parameters of the scene where the zebra crossing is located, wherein the plurality of scene parameters of the scene where the zebra crossing is located are a zebra crossing distribution length, a zebra crossing distribution area, and an area occupied by a traffic intersection where the zebra crossing is located in the scene where the zebra crossing is located; A customized recognition mechanism is connected to the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively, and is used to synchronously input multiple scene parameters of the scene where the zebra crossing is located, various imaging depths of field corresponding to various edge pixel points of the main object in the current shooting picture, various coordinate values, and multiple working parameters of the photoelectric sensor component of the scene shooting component into the AI recognition model, and execute the AI recognition model to obtain the type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model; Among them, each imaging depth of field corresponding to each edge pixel point of the main object in the current shooting picture, each coordinate value, multiple working parameters of the photoelectric sensor component of the scene shooting component, and the number of noise types of the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The multiple binary numerical values corresponding to the distribution of multiple image processing types executed in the scene shooting component include: the multiple working parameters of the photoelectric sensor component of the scene shooting component are the electronic shutter rate, frame rate, signal-to-noise ratio and quantum efficiency of the photoelectric sensor component of the scene shooting component; Among them, the imaging depths of field, the coordinate values, the multiple working parameters of the photoelectric sensor components of the scene shooting components and the number of noise types in the current shooting picture corresponding to the edge pixel points of the main object in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the output of the current shooting picture. The type names represented by multiple binary values corresponding to the distribution of multiple image processing types executed in the scene shooting component also include: the main object in the current shooting picture is the object occupying the largest area in the current shooting picture.
2. The zebra crossing scene shooting picture recognition system according to claim 1, characterized in that: The imaging depths of field, coordinate values, multiple working parameters of the photoelectric sensor component of the scene shooting component and the number of noise types of the current shooting picture corresponding to each edge pixel point of the main object in the current shooting picture are synchronously input into the AI recognition model, and the AI recognition model is executed to obtain the type names of the multiple binary numerical representations corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model, which also include: the type names of the multiple binary numerical representations corresponding to the distribution of multiple image processing types executed in the scene shooting component of the current shooting picture output by the AI recognition model are the representation mode of multiple binary numerical values connected end to end.
3. The zebra crossing scene shooting picture recognition system according to claim 2, characterized in that: The system further comprises: a cache execution device, which is arranged near the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is connected to the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively; Among them, the cache execution device is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the cache execution device is used to respectively implement the cache operation of the input signals and output signals of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component.
4. The zebra crossing scene shooting picture recognition system according to claim 2, characterized in that: The system further comprises: A positioning server component is arranged near the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is connected to the custom recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively; Among them, the positioning server component is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the positioning server component is used to provide positioning data services based on the Beidou navigation mechanism for the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component respectively.
5. The zebra crossing scene shooting image recognition system according to any one of claims 2 to 4, characterized in that: A timing server component is used which is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, and is used to provide the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component with their respective required timing services.
6. The zebra crossing scene shooting picture recognition system according to claim 5, characterized in that: A timing server device is provided near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component for providing the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component with their respective required timing services, including: the timing server device is an FPGA device.
7. The zebra crossing scene shooting image recognition system according to any one of claims 2 to 4, characterized in that: A serial port configuration device is arranged near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, for providing on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component.
8. The zebra crossing scene shooting image recognition system according to claim 7, characterized in that: A serial port configuration device is provided near the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component for providing on-site configuration operations of working parameters based on serial port connection for each of the customized recognition mechanism, the parameter analysis mechanism, the scene shooting component and the continuous learning component, including: the serial port configuration device is a programmable logic device.
9. The zebra crossing scene shooting picture recognition system according to claim 8, characterized in that: A serial port configuration device is used which is arranged near the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component and is respectively connected to the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component to provide on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component. The serial port configuration device also includes: the serial port configuration device uses different configuration addresses to provide on-site configuration operations of working parameters based on the serial port connection for each of the customized recognition mechanism, the parameter parsing mechanism, the scene shooting component and the continuous learning component.
Citation Information
Cited By
Visual data learning system for human-computer interaction
CN119987560A
