Instant intention unlocking system based on facial recognition and voice interaction

Abstract

The invention discloses an instant intention unlocking system based on facial recognition and voice interaction, and aims at solving the problems that an existing voice assistant depends on wake-up words, interaction is low in efficiency, scene perception is weak, and energy consumption is high. The system adopts a three-layer architecture including a hardware layer (a camera, a microphone, a processor and a storage module), a software layer (a voice recognition engine, a face recognition algorithm, a scene perception module and an authority control module) and a logic layer (an intention fusion judgment module and a task routing and execution module). 'gaze + voice 'is taken as core triggering, and wakeup words are avoided; face recognition is used for confirming identity and watching state, voice recognition is used for analyzing instructions, and an application scene is fused to generate an instant intention triggering model; authority hierarchical management and control are realized, and energy consumption optimization, auxiliary triggering, cross-terminal synchronization and emergency help modules are arranged. Interaction efficiency can be improved, safety and convenience are balanced, energy consumption is reduced, and the method is adaptive to multiple terminals and multiple scenes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human-computer interaction and intelligent terminal control, and particularly relates to an instant intention unlocking system based on facial recognition and voice interaction. BACKGROUND

[0002] With the popularization of intelligent terminal devices and the rapid development of artificial intelligence technology, voice interaction has become an important way of human-computer interaction. The mainstream voice assistant products on the market, such as Siri, Xiaoai, Alexa, etc., generally use a trigger mechanism based on a wake-up word. The user needs to say a specific wake-up word, such as "Hey Siri", "Xiaoai", etc., to activate the voice assistant to enter a listening state, and then issue a specific operation instruction. Although this interaction mode realizes voice control to some extent, it still has many shortcomings.

[0003] The existing technology has exposed four main problems in actual application: first, the operation efficiency is low, the wake-up word trigger needs the user to actively say a specified password, there is obvious delay in the whole interaction process, the user experience is fragmented, and real-time response cannot be realized; second, the system island problem is prominent, the existing facial recognition technology is mostly used for single device unlocking function, and cannot be deeply integrated with the intention recognition system of the voice assistant, lacking the organic integration of natural language intention and identity confirmation; third, lacking scene perception ability, the voice assistant usually runs as an independent module, and it is difficult to automatically match the user's intention combined with the context environment of the current foreground application, resulting in low intelligentization of interaction; fourth, the energy consumption problem is serious, the voice assistant needs to reside in the background to continuously listen to the wake-up word, causing high device power consumption, which seriously affects the endurance of mobile devices.

[0004] In summary of the above analysis, there is currently no mature technical solution for deeply integrating facial recognition identity verification and voice instant intention routing. The existing technology cannot realize the integrated interaction mode of "natural gaze + voice intention", and it is difficult to meet the user's demand for natural, efficient and intelligent human-computer interaction experience. Therefore, it is urgent to develop a new type of interaction system that can get rid of the limitation of wake-up word, realize instant voice interaction based on gaze trigger, and provide personalized and secure services according to user identity and use scenarios. SUMMARY

[0005] This invention aims to solve the technical problems of existing voice assistants, such as reliance on wake words, limited facial recognition functionality, low interaction efficiency, and lack of scene awareness. Specifically, this invention addresses the following technical challenges: how to achieve natural voice interaction without wake words; how to deeply integrate facial recognition with voice intent recognition; how to intelligently route user intents based on application scenarios; how to reduce system power consumption while ensuring real-time interaction; how to provide differentiated permission management in multi-user environments; how to ensure the security of high-risk operations; and how to achieve cross-terminal intent synchronization.

[0006] To address the aforementioned technical problems, this invention provides an instant intent unlocking system based on facial recognition and voice interaction, comprising: The hardware layer includes the camera, microphone, processor, and storage module; The software layer includes a speech recognition engine, a facial recognition algorithm, a scene awareness module, and an access control module; The logic layer includes an intent fusion and determination module and a task routing and execution module; The camera is used to capture images of the user's face and detect their gaze. The microphone is used to acquire a short-term speech signal after a gaze is detected; Facial recognition algorithms are used to identify users and confirm their gaze status; A speech recognition engine is used to convert speech signals into text information; The scene awareness module is used to obtain the current status of active foreground applications or background services; The intent fusion and determination module is used to fuse user identity information, voice parsing results, and application status to generate an instant intent triggering model; The task routing and execution module triggers the model to unlock and execute the corresponding task based on the real-time intent.

[0007] Furthermore, the access control module includes: The intent classification unit is used to classify task intents into normal operations or high-risk operations. The direct execution unit is used to directly allow the execution of ordinary operations. The secondary confirmation unit is used to generate confirmation prompts for high-risk operations and wait for secondary confirmation from the user. High-risk operations include payment transfers, data deletion, and system unlocking.

[0008] Furthermore, the system includes an energy consumption optimization mechanism: The camera uses a low-power continuous monitoring mode, performing only basic gaze detection; The speech recognition engine is in a dormant state when not looking at the viewer and is only activated when a gaze is detected. The voice recognition engine sets a short listening window of 5-10 seconds; The processor dynamically adjusts the working frequency according to the system state.

[0009] Further, the system further comprises: A user identity database storing facial feature information of different users and corresponding permission configurations; After the facial recognition algorithm identifies the user's identity, the system automatically loads the user's corresponding permission settings and personalized preferences.

[0010] Further, the system further comprises an auxiliary trigger extension module, which comprises at least one of the following: A lip reading unit for assisting voice recognition in a noisy environment; An iris recognition unit or a voiceprint recognition unit for assisting identity confirmation when the face recognition accuracy decreases; A pressure sensor for detecting the user's holding state in a driving scenario; Wherein the gaze and voice are the core trigger conditions, and the auxiliary trigger is only used as an enhancement measure.

[0011] Further, the system further comprises an intent synchronization service module: The intent synchronization service module runs in the cloud or a local network; When a terminal device generates an instant intent trigger model, the intent synchronization service module synchronizes the model to other terminal devices of the user through encrypted transmission; The terminal device includes a mobile phone, a tablet, a vehicle-mounted system, AR glasses, and a smart speaker.

[0012] Further, the secondary confirmation unit comprises: A speech synthesizer for generating a voice confirmation prompt; A confirmation receiver for receiving the user's voice reply or screen click confirmation; A threshold determiner for requiring additional biometric verification when the payment amount exceeds a preset threshold.

[0013] Further, the system further comprises an emergency help module: A keyword recognition unit for recognizing emergency keywords such as "uncomfortable" and "help"; An emergency plan execution unit for skipping the permission confirmation process, automatically dialing an emergency number, sending location information to emergency contacts, turning on the flashlight and alarm sound after recognizing the emergency keyword.

[0014] Further, the identity recognition and gaze confirmation of the facial recognition algorithm are completed within 100 milliseconds; The voice recognition of the voice recognition engine is completed within 500 milliseconds; The overall response time from gaze detection to task execution is within 2 seconds.

[0015] The above-mentioned instant intent unlocking system based on face recognition and voice interaction adopts a three-layer architecture design of hardware layer, software layer and logic layer. The hardware layer includes a face recognition camera, a voice collection microphone, a processor and a storage module, which provides basic data collection and processing capability for the system. The software layer includes a voice recognition engine, a face recognition algorithm, a scene perception module and a permission control module, which realizes intelligent analysis and processing of collected data. The logic layer includes an intent fusion judgment module and a task routing and execution module, which is responsible for generating an instant intent trigger model and executing corresponding tasks.

[0016] The core innovation of the present application lies in the establishment of eight key technical mechanisms: joint intent trigger model, identity and gaze state confirmed by face recognition, unique instant intent trigger model generated by combining voice analysis; wake-up word free design, gaze as trigger signal, short-time voice monitoring of device activated, user can execute instructions naturally; scene perception and intent routing, combined with foreground active application or background service, automatically matching instruction target; permission grading and secondary confirmation, secondary semantic confirmation for high-risk operations such as payment, deletion and unlocking; multi-user differentiation mechanism, automatically loading different user permissions or preferences based on face recognition; energy optimization mechanism, short-time voice monitoring only when gaze triggers, significantly reducing power consumption; auxiliary trigger expansion mechanism, combining other biological characteristics or sensing signals as auxiliary confirmation conditions to enhance system robustness; multi-terminal collaboration, supporting seamless intent synchronization between mobile phones, tablets, vehicles, AR, sound boxes and other terminals.

[0017] The beneficial effects of the present application are: Compared with the prior art, the present application has the following significant advantages: 1. Natural interaction is realized through the "gaze + voice" dual trigger mechanism, which completely gets rid of the shackles of wake-up words, simplifies the traditional "wake-up-wait-command" three-step operation to "gaze-command" two steps, significantly improves the interaction efficiency, and makes the human-computer interaction more natural and smooth; the deep integration of face recognition and voice intent is realized, which integrates identity confirmation, permission verification, intent recognition and task execution into an integrated process, breaking the situation that each functional module is independent in traditional technology, and providing a more intelligent and secure interaction experience.

[0018] 2. The scene perception routing mechanism is introduced, the system can automatically understand the target object of user intent according to the currently running application program, which improves the intelligent level and context adaptation ability of the system, and reduces the amount of information the user needs to provide; a perfect hierarchical permission confirmation mechanism is established, which ensures instant response for ordinary operations while implementing secondary confirmation for high-risk operations, effectively balancing the needs of convenience and security.

[0019] 3. Employing innovative energy-saving optimization solutions, through short-term monitoring windows and dynamic frequency adjustment, it can reduce standby power consumption by more than 80% compared to traditional resident monitoring solutions, greatly improving the battery life of mobile devices; it supports multi-user identity differentiation and personalized configuration, and in shared device scenarios such as home and office, it can automatically identify different users and load corresponding permissions, protecting privacy and providing a personalized experience; it realizes cross-terminal intent synchronization, allowing users to seamlessly continue previous tasks when switching between different devices, truly achieving a user-centric continuous experience and expanding the system's application scope and usage scenarios. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 The overall system framework of this invention is shown in the three-layer architecture (hardware layer, software layer, and logic layer) and the main modules contained in each layer. Figure 2 The interactive flowchart of this invention illustrates the six-step process from user gaze to task execution. Figure 3 The flowchart for the permission level confirmation of this invention illustrates the hierarchical processing mechanism where ordinary operations are executed directly, while high-risk operations require secondary confirmation. Figure 4 This is a schematic diagram of a typical application scenario of the present invention (taking a picture): It shows a scenario where the camera application is automatically launched after the user looks at the phone and says "take a picture"; Figure 5 This is a schematic diagram of a multi-user differentiated scenario (in-vehicle device) for the present invention: illustrating different permission configurations when a father and a child use the same in-vehicle device; Figure 6 This is a schematic diagram of cross-terminal synchronization according to the present invention: it shows the process of synchronizing the intention generated by the mobile phone to the vehicle device for continued execution. Detailed Implementation

[0022] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0023] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0024] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0025] See Figures 1 to 6 As shown This invention provides an instant intent unlocking system based on facial recognition and voice interaction, as shown in the attached figure. Figure 1 As shown, the system adopts a three-layer architecture, including hardware layer 3, software layer 2, and logic layer 1. Hardware layer 3 includes a camera for capturing user facial images, a microphone for capturing voice signals, a processor providing computing power, and a storage module for data storage. Software layer 2 includes a speech recognition engine for converting captured voice signals into text information; a facial recognition algorithm for identifying the user and detecting gaze state; a scene awareness module for obtaining the current status of active foreground applications or background services; and an access control module for managing permissions for different types of operations. Logic layer 1 includes an intent fusion and determination module for fusing facial recognition results with speech parsing results to generate an instant intent triggering model; and a task routing and execution module for unlocking and executing corresponding tasks based on the instant intent triggering model.

[0026] As attached Figure 2As shown, the system's workflow is as follows: First, the camera continuously detects whether the user is looking at the device. When the user's gaze is detected, the facial recognition algorithm simultaneously performs identity verification and gaze confirmation. Once the user is confirmed to be looking, the system immediately activates a short-term voice monitoring function, and the microphone begins to collect voice signals, waiting for the user to input voice commands. The user does not need to say a wake word and can directly express the task intent through natural language, such as "take a picture" or "navigate to the company." The voice recognition engine recognizes the collected voice signals and parses the specific task intent through the semantic parsing module. The intent fusion and determination module fuses the user identity information obtained from facial recognition with the task intent obtained from voice parsing, and combines it with the current application state obtained by the scene perception module to generate a unique instant intent trigger model. The system determines the legality of the task based on this trigger model. If it is a legal task, the unlocking operation is completed, and the corresponding function is executed through the task routing and execution module.

[0027] As attached Figure 3 As shown, for high-risk operations, the system employs a tiered access control mechanism. After the intent fusion and judgment module generates a task intent, the access control module first categorizes and determines whether it is a normal or high-risk operation. For normal operations such as opening the camera, playing music, or adjusting volume, the system directly executes the corresponding function, ensuring the immediacy of the operation. For high-risk operations such as making payments, deleting important data, or unlocking the system, the access control module triggers a secondary confirmation process. The system requires the user to confirm again via voice synthesis or on-screen prompts, for example, "You are about to transfer 1000 yuan to Zhang San, please confirm again." The user needs to reply "confirm" via voice or other preset confirmation methods before the system executes the high-risk operation, thus ensuring system security.

[0028] The system's energy consumption optimization mechanism is achieved through hardware and software collaboration. At the hardware layer (3 sides), the camera employs a low-power continuous monitoring mode, performing only basic gaze detection without complete facial recognition calculations. At the software layer (2 sides), the speech recognition engine is normally in sleep mode, activating only upon detecting gaze, and a short listening window (e.g., 5-10 seconds) is set. If no valid voice command is received within this time, the system automatically disables the voice listening function and returns to standby mode. The processor dynamically adjusts its operating frequency based on the system state, reducing the frequency to save energy during standby and increasing the frequency to ensure processing speed when speech or facial recognition is required.

[0029] As attached Figure 5As shown, this system supports multi-user differentiation. The facial recognition algorithm maintains a user identity database, storing facial feature information and corresponding permission configurations for different users. When the system identifies a different user, it automatically loads that user's permission settings and personalized preferences. Taking an in-vehicle device as an example, when the father looks at the in-vehicle screen and says "navigate home," the system identifies his identity, loads full permissions, allows access to the navigation function, and automatically calls his preset home address for route planning. When the child looks at the same device, the system identifies his restricted identity, only granting access to entertainment functions and blocking navigation, payment, and other functions. Even if the child gives related instructions, the system will prompt "You do not have permission to use this function."

[0030] The system's auxiliary triggering extension mechanism allows for the integration of other biometric features or sensor signals to enhance the accuracy of intent verification. For example, in noisy environments, the system can combine lip reading technology to assist speech recognition; when facial recognition accuracy decreases due to users wearing masks, iris recognition or voiceprint recognition can be combined for identity verification; in driving scenarios, the system can use a pressure sensor on the steering wheel to detect the user's grip as an auxiliary verification condition. However, these auxiliary methods are only enhancement measures; gaze and voice remain the core triggering conditions, and the system does not rely on gesture operations as the primary triggering method.

[0031] As attached Figure 6 As shown, this system supports cross-terminal real-time intent synchronization. The system maintains an intent synchronization service in the cloud or on a local network. When a user generates a real-time intent trigger model on one terminal device, this model can be synchronized to the user's other terminal devices via encrypted transmission. For example, if a user looks at and says "Navigate to the company" on their phone, the phone completes intent recognition and synchronizes the intent to the user's in-vehicle device. When the user enters the car, the in-vehicle device automatically displays a prompt asking "Continue navigation to the company?". After the user confirms, the navigation task can be seamlessly continued. The synchronization mechanism supports multiple terminal devices such as mobile phones, tablets, in-vehicle systems, AR glasses, and smart speakers, ensuring a continuous user experience in different scenarios.

[0032] Example 1 As attached Figure 4The following diagram illustrates the system's implementation process using a user taking a photo with their mobile phone. When a user holds their smartphone and wants to take a picture, they look at the screen. The front-facing camera detects this gaze, and the facial recognition algorithm completes user identification and gaze confirmation within 100 milliseconds. The system immediately activates the microphone for brief voice monitoring, and the user directly says "take a picture" without any wake-up word. The voice recognition engine completes voice recognition within 500 milliseconds, deciphering the "take a picture" intention. The intention fusion and determination module combines the user's identity information with the intention to take a picture, generating an instant intention trigger model. Since taking a picture is a normal operation, the access control module directly allows it. The system unlocks, the task execution module calls the camera API, launches the camera application, and enters the photo-taking interface. The entire process is completed within 2 seconds, achieving a natural and fast interactive experience.

[0033] Example 2 Application in a medical setting: Doctors wear AR glasses during surgery. When they need to check a patient's vital signs, they look at the display area of ​​the AR glasses. The built-in miniature camera detects the doctor's gaze and confirms their identity. The system then activates voice monitoring; the doctor says "display blood pressure," and the voice recognition engine recognizes the command and interprets the intent. Because medical data is sensitive information, the access control module first verifies the doctor's professional qualifications and the current surgical authorization. After confirmation, the system calls the medical information system interface to obtain the patient's real-time blood pressure data and displays "Blood Pressure: 120 / 80 mmHg" in a semi-transparent manner within the AR glasses' field of view. The entire process does not interfere with the doctor's surgical operations, achieving efficient information acquisition without hands-free operation.

[0034] Example 3 Application Implementation in a Home Setting. A user is lying on the sofa watching TV and wants to fast-forward the current video. The user looks at the TV screen, and the TV's built-in camera detects this gaze and confirms the user is a family member through facial recognition. The system then activates voice monitoring. The user says, "Fast forward 20 seconds." After voice recognition, the scene perception module detects that the current foreground application is a video player, and the intent fusion and determination module generates a fast-forward command for the video player. Since it's a normal media control operation, the system executes it directly, and the video automatically fast-forwards by 20 seconds. If the user continues by saying, "Turn the volume up," the system doesn't need to re-detect the gaze; it directly recognizes and executes the volume adjustment within a short window, achieving continuous and natural interaction.

[0035] Example 4 Application Implementation in Mobile Payment Scenarios. A user is shopping in a mall, opens Alipay to pay, and looks at the phone screen. The system recognizes the user's identity and detects their gaze. The user says "Scan to pay," and after the voice recognition engine analyzes the intent, the access control module determines that the operation is a high-risk financial transaction, triggering a secondary confirmation process. The system prompts via voice synthesis, "You are about to make a scan payment, please confirm again," while a confirmation button is displayed on the screen. The user can reply "Confirm payment" via voice or click the button on the screen. The system only opens the scanning interface after receiving secondary confirmation. If the payment amount exceeds a preset threshold, such as 1000 yuan, the system will also require additional biometric verification, such as fingerprint or more precise facial recognition, to ensure payment security.

[0036] Example 5 Application Implementation in Emergency Situations. An elderly user suddenly feels unwell. Due to weakness, they are unable to perform complex operations. The user strains to focus on their smartphone placed on the bedside table. Upon detecting this focused gaze, the system initiates voice monitoring. The user weakly utters "I'm not feeling well" or "Help!" The voice recognition engine identifies the emergency keywords, and the intent fusion and judgment module immediately generates an emergency assistance intent. The system skips the conventional permission confirmation process and directly executes the emergency plan: automatically dialing 120 (emergency services), simultaneously sending location and assistance information to pre-set emergency contacts, and activating the phone's flashlight and alarm to attract attention. The entire process is fully automated, maximizing the user's ability to receive timely assistance in emergency situations.

[0037] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0038] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A real-time intent unlocking system based on facial recognition and voice interaction, characterized in that, include: The hardware layer includes the camera, microphone, processor, and storage module; The software layer includes a speech recognition engine, a facial recognition algorithm, a scene awareness module, and an access control module; The logic layer includes an intent fusion and determination module and a task routing and execution module; The camera is used to capture images of the user's face and detect gaze status. The microphone is used to acquire short-term speech signals after a gaze is detected; The facial recognition algorithm is used to identify the user's identity and confirm their gaze state; The speech recognition engine is used to convert speech signals into text information; The scene perception module is used to obtain the current status of active foreground applications or background services. The intent fusion and determination module is used to fuse user identity information, voice parsing results and application status to generate an instant intent triggering model. The task routing and execution module completes the unlocking and executes the corresponding task based on the real-time intent trigger model.

2. The system according to claim 1, characterized in that, The access control module includes: The intent classification unit is used to classify task intents into normal operations or high-risk operations. The direct execution unit is used to directly allow the execution of ordinary operations. The secondary confirmation unit is used to generate confirmation prompts for high-risk operations and wait for secondary confirmation from the user. The high-risk operations mentioned above include payment transfers, data deletion, and system unlocking.

3. The system according to claim 1, characterized in that, The system includes an energy consumption optimization mechanism: The camera uses a low-power continuous monitoring mode and only performs basic gaze detection. The speech recognition engine is in a dormant state when not looking at the view, and is only activated when a gaze is detected. The speech recognition engine is set to a short listening window of 5-10 seconds; The processor dynamically adjusts its operating frequency based on the system status.

4. The system according to claim 1, characterized in that, The system also includes: The user identity database stores facial feature information and corresponding permission configurations for different users. After the facial recognition algorithm identifies the user, the system automatically loads the user's corresponding permission settings and personalized preferences.

5. The system according to claim 1, characterized in that, The system further includes an auxiliary triggering expansion module, which includes at least one of the following: Lip reading unit for assisting speech recognition in noisy environments; Iris recognition units or voiceprint recognition units are used to assist in identity verification when facial recognition accuracy declines. Pressure sensors are used to detect the user's grip status in driving scenarios; Gaze and speech are the core triggering conditions, while auxiliary triggers are only used as enhancement measures.

6. The system according to claim 1, characterized in that, The system also includes an intent synchronization service module: The intent synchronization service module runs in the cloud or on a local network; Once a terminal device generates an instant intent trigger model, the intent synchronization service module synchronizes the model to the user's other terminal devices via encrypted transmission. The terminal devices include mobile phones, tablets, in-vehicle systems, AR glasses, and smart speakers.

7. The system according to claim 2, characterized in that, The secondary confirmation unit includes: A speech synthesizer is used to generate voice confirmation prompts; Confirmation receiver, used to receive user's voice reply or screen tap confirmation; A threshold checker is used to require additional biometric verification when the payment amount exceeds a preset threshold.

8. The system according to claim 1, characterized in that, The system also includes an emergency assistance module: Keyword recognition unit, used to identify urgent keywords; The emergency response execution unit is used to skip the authorization confirmation process after recognizing emergency keywords, automatically dial emergency numbers, send location information to emergency contacts, and activate flashing lights and alarms.

9. The system according to claim 1, characterized in that: The facial recognition algorithm completes identity recognition and gaze confirmation within 100 milliseconds; The speech recognition engine completes speech recognition within 500 milliseconds; The system's overall response time from gaze detection to task execution is within 2 seconds.

Images