Oral muscle training system

Abstract

The present invention discloses an oral muscle training system, comprising an oral muscle training space and an oral muscle training method on a computer built into the oral muscle training space, wherein a video module is installed on the top of the inner layer of the oral muscle training space, and a temperature and humidity control module is installed on the front and back lower parts of the inner layer of the oral muscle training space. The outer layer of the oral muscle training space is a solid wall, and the inner layer of the oral muscle training space is composed of 5 LED display screens. The video module, temperature and humidity control module, and LED display screen are all connected to the computer by wired or wireless communication; the oral muscle training method comprises the following steps: independently designing a training script or importing a standard training script; completing the training according to the training script; and evaluating the user to obtain an evaluation result. The present invention can provide a more interesting and effective training method for patients with oral muscle disorders, combining education with entertainment. The system of the present invention can automatically adjust the temperature and humidity during the training process to help users reduce sensitivity.

Description

Technical Field

[0001] The present invention relates to the technical field of oral muscle rehabilitation, in particular to the technical field of oral muscle training systems. Background Art

[0002] Rehabilitation medicine is a comprehensive discipline that applies a variety of effective measures to alleviate and compensate for the physical and mental impairments of injured, ill, and disabled individuals, maximizing their remaining functions and enabling them to return to their families and participate in society in optimal condition. Building on disciplines such as orthopedics, physics, neuroscience, and sports medicine, a diverse range of rehabilitation therapies and theories have flourished, including physical therapy, occupational therapy, speech-language therapy, and psychological counseling and analysis. Recently popular brain-computer interfaces, digital therapies, and play therapy are also rising stars in rehabilitation medicine. Rehabilitation for oral muscle disorders (hereafter referred to as "oral muscles") is a niche area within this field.

[0003] A variety of disorders may be associated with oral muscle problems, such as autism, Down syndrome, language delay, articulation disorders, fluency disorders, voice problems, and swallowing disorders. For patients with oral muscle disorders without organic lesions or physiological disorders, traditional oral muscle intervention methods mainly rely on diagnosis and assessment by hospitals or professional therapists, followed by continuous oral muscle training with the help of rehabilitation therapists. This method requires a relatively long rehabilitation process and is highly dependent on the professional level of rehabilitation therapists. After leaving the institution, patients often lack home rehabilitation methods, making it difficult to persist in the long term. Summary of the Invention

[0004] The present invention aims to address the challenges of the existing technology by providing an oral muscle training system that offers a more engaging and effective training method for patients with oral muscle disorders, combining education with entertainment. Some oral muscle disorders are caused by excessive sensitivity of the skin or other parts of the body to liquids or temperature, resulting in impairments in oral muscle movements. The system of the present invention can automatically adjust temperature and humidity during training to help users reduce sensitivity.

[0005] To achieve the above objectives, the present invention provides an oral muscle training system, comprising an oral muscle training space and an oral muscle training method on a computer built into the oral muscle training space. A video module is installed on the top of the inner layer of the oral muscle training space, and a temperature and humidity control module is installed on the front and back lower portions of the inner layer of the oral muscle training space. The outer layer of the oral muscle training space is a solid wall, and the inner layer of the oral muscle training space is composed of five LED screens. The video module, temperature and humidity control module, and LED screen are all connected to the computer by wired or wireless communication.

[0006] The oral muscle training method comprises the following steps:

[0007] Step 1: Design your own training script or import a standard training script;

[0008] Step 2: Complete training according to the training script;

[0009] Step 3: Evaluate the user and obtain the evaluation results.

[0010] Preferably, the computer has a data processing hub with the ImageBind large model as its core.

[0011] Preferably, the video module includes an ultra-short-throw projector and a camera. The ultra-short-throw projector is used to make the inner floor become a screen, and the camera is used to collect the video and audio of the user during the training process.

[0012] Preferably, the temperature and humidity control module is used to control temperature and humidity, and perform heating or cooling, humidification or drying.

[0013] Preferably, the inner layer includes 5 LED display screens, the 4 walls and the top of the inner layer are all LED display screens, and there is a gap between the inner layer and the outer layer and the gap is between 50-100 cm.

[0014] Preferably, the inner floor is provided with a drainage device.

[0015] Preferably, the self-designed training script of step 1 is specifically:

[0016] Step z11: Select training mode and training goal: the training mode includes confidence mode and challenge mode;

[0017] Step z12: Input training script generation requirements and generate a training script: The user inputs training script generation requirements, and the AI ​​center in the computer generates a training script based on the requirements. The requirements include text descriptions, pictures, audio, and video. The training script includes elements such as scene generation, temperature and humidity control during script training, oral muscle training tutorials, music matching, and determination of confidence encouragement mechanisms or challenge assessment mechanisms.

[0018] Preferably, the importing standard training script in step 1 is specifically as follows:

[0019] Step d11: Select a training mode: the training mode includes a confidence mode and a challenge mode;

[0020] Step d12: Output the pre-imported standard training script: The standard training script includes the following elements: scene generation, temperature and humidity control during script training, oral muscle training tutorial, music matching, and determination of confidence encouragement mechanism or challenge assessment mechanism.

[0021] Preferably, a debate encouraging system is further included, wherein the debate encouraging system is used to extract specific segments of audio and video during the user training process, and then ask questions to the user to encourage the user to debate.

[0022] Beneficial effects of the present invention:

[0023] 1. The present invention provides a more interesting and effective training method for patients with oral muscle disorders, which combines education with entertainment.

[0024] 2. The cause of oral muscle disorders in some patients is that their skin or other parts are overly sensitive to liquids or temperature, resulting in obstruction of oral muscle-related movements. The system of the present invention can automatically adjust the temperature and humidity during training to help users reduce sensitivity.

[0025] 3. With the help of artificial intelligence technology, the system of the present invention allows users to independently design training scripts or import standard training scripts. The training modes are optional and the scenes are changeable. At the same time, the music of the training script has corresponding rhythmic syllables to match, helping users to train better.

[0026] 4. After the system of the present invention is trained, it does not simply output an annoying score number, but is divided into "confidence mode" and "challenge mode" evaluation. At the same time, it introduces an "encourage debate system" to make evaluation a form of training.

[0027] The features and advantages of the present invention will be described in detail through embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a hardware rendering of an oral muscle training system according to the present invention;

[0029] Figure 2 It is a system logic diagram of an oral muscle training system of the present invention;

[0030] Figure 3 It is a technical principle diagram of an oral muscle training system of the present invention.

[0031] In the figure: 1-video module, 2-temperature and humidity control module, 3-outer layer, 4-inner layer. DETAILED DESCRIPTION

[0032] See Figure 1 、 Figure 2 、 Figure 3The present invention comprises an oral muscle training space and an oral muscle training method on a computer built in the oral muscle training space, wherein a video module (1) is installed on the top of an inner layer (4) of the oral muscle training space, a temperature and humidity control module (2) is installed on the front and rear lower parts of the inner layer (4) of the oral muscle training space, an outer layer (3) of the oral muscle training space is a solid wall, and the inner layer (4) of the oral muscle training space is composed of 5 LED display screens, and the video module (1), the temperature and humidity control module (2), and the LED display screen are all connected to the computer by wired or wireless communication;

[0033] The oral muscle training method comprises the following steps:

[0034] Step 1: Design your own training script or import a standard training script;

[0035] Step 2: Complete training according to the training script;

[0036] Step 3: Evaluate the user and obtain the evaluation results.

[0037] The present invention also includes a debate encouraging system, which is used to extract specific segments of audio and video during the user training process and then ask questions to the user to encourage the user to debate.

[0038] Working process of the present invention:

[0039] The oral muscle training system of the present invention is described in conjunction with the accompanying drawings during operation.

[0040] Benefiting from the rapid development of artificial intelligence technology, this system was developed in the hope that it can provide patients with oral muscle disorders with more training methods and make learning fun.

[0041] This system utilizes numerous AI-related technologies, with its data processing hub being based on the ImageBind AI model. ImageBind, a large open-source model released by Meta in May 2023, embeds text, audio, visual, thermal (infrared), depth, and IMU data (inertial measurement unit, which measures and reports acceleration, orientation, angular rate, and other gravity-related data) into a vector space, orchestrating and processing these six modalities. It not only interacts with users through everyday chat but also generates or processes existing data from these six modalities to generate analysis results tailored to their specific needs. For example, a user can upload an image and type in the query, "What animal is in this image? Where is it located?" ImageBind will automatically analyze the image elements and conduct online analysis, ultimately outputting the result, "The animal in the image is a Bengal tiger, located at a zoo in the southern United States." ImageBind is the data processing hub of this system, responsible for overall data collection, processing, and summarization.

[0042] In this system, users can customize the generation of virtual characters to accompany them during training. This technology relies on technologies such as Shap-E and OpenAvatar. Shap-E is a conditional generative model developed specifically for 3D character design, and OpenAvatar is an open-source digital human avatar framework that provides a set of basic movements and cross-platform support. Furthermore, this system collects video and audio data through visual capture using a camera. This process relies on the DeepLabCut algorithm toolkit, which has been widely used for automated behavioral analysis of humans, mice, and macaques. This toolkit automatically labels different parts of an object (such as different parts of the human body) in a video and extracts the motion trajectory of each part with millisecond-level temporal accuracy. Based on this information, dynamic time alignment kernelization algorithms and unsupervised learning algorithms are then used to achieve automated information classification and output quantitative data reports.

[0043] Users need to carry out training in the oral muscle training space. The specific settings of the oral muscle training space are as follows: Figure 1 As shown. The video module 1 on the top of the inner layer 4 consists of two parts: an ultra-short-throw projector and a camera. The video module 1 can make the floor of the inner layer 4 a screen, presenting various images, and at the same time collect videos and audios of the user during training. The temperature control and humidity control module 2 controls the temperature, heating or cooling (16°C-35°C) and humidifying or drying. The space is divided into two layers, the outer layer 3 is an ordinary room with solid walls on all sides, and the inner layer 4 consists of 5 LED screens (4 screens make up the 4 walls, 1 screen makes up the roof of the inner layer 4, there is no screen on the ground, and the projector is installed in the center of the roof); there is a gap between the walls of the inner and outer layers, and the four walls of the inner layer 4 (4 screens) can move freely back and forth, and the gap between them and the walls can be adjusted between 50-100cm. The ground of the inner layer 4 is equipped with a drainage device.

[0044] When using this system, users first need to design their own training scripts or import standard training scripts, then complete the training and view the evaluation results. Figure 2 .

[0045] Based on the artificial intelligence data processing center, the self-designed training script is divided into the following steps:

[0046] Select training mode and training target:

[0047] There are two training modes: confidence mode and challenge mode.

[0048] Confidence mode: The overall design of the script is mainly guided by encouragement and confidence enhancement. It does not set assessment or evaluation measures such as scoring and rating. It highlights the advantages and highlights of the training process and displays them in a centralized manner to enhance user confidence.

[0049] Challenge Mode: Specific training goals need to be set, and the Artificial Intelligence Data Center (AI Center) will output scores based on multiple objective indicators to assess whether the user has achieved the goal and completed the challenge.

[0050] For example, the user selects "confidence mode" and sets the training goal to "persist in training for 5 minutes"; the user selects "challenge mode" and sets the training goal to "complete the script task and score more than 90 points (out of 100)".

[0051] Enter the training script generation requirements and generate the training script:

[0052] The user needs to input the script generation requirements, and the AI ​​center will generate a specific training script based on the requirements. This system allows users to submit their own script generation requirements in 4 ways:

[0053] 1. Text description: For example, if the user enters "beach and waves, breeze, suitable temperature, I hope to train at sunset, and there will be some marine animals to accompany me in training", the AI ​​center will automatically generate a training script based on this.

[0054] 2. Pictures: Users can upload their favorite pictures as script generation requirements. For example, if they upload a picture of a beach, the AI ​​center will generate the corresponding script based on "beach".

[0055] 3. Audio: Users can upload an audio clip as a script generation requirement, for example, the sound of waves. The AI ​​hub will then include and apply the “sound of waves” in the generated script.

[0056] 4. Video: Users can upload a video as a script generation requirement. The video length should not exceed 5 minutes. For example, if you upload a video of "taking wedding photos on the beach with children running around", the AI ​​center will use this as a benchmark to automatically extract keywords of video elements and generate a script of "beach, wedding dress, children, and weather elements in the video."

[0057] These four methods are not isolated but rather coordinated with each other. Users can first describe their requirements in text and then add images, audio, or video. They can also upload an image and then use other methods to assist in describing the requirements. The coordination of these four methods can make the AI-generated training scripts more accurate and meet user needs.

[0058] The script generated by the AI ​​Hub contains the following elements:

[0059] 1. Scene generation: The scenes in the script contain two categories (cartoon and real scene). Each category of scenes will be presented in the inner layer (such as Figure 1)’s four walls, roof, and ground, creating a virtual naked-eye 3D scene. After the scene is generated, users can freely choose whether to include the “protagonist” in the scene.

[0060] a. Protagonist Appearance: Integrating technologies such as Shap-E and OpenAvatar, the AI ​​Hub allows users to customize and generate a unique virtual protagonist through the four methods mentioned above. This protagonist then becomes the protagonist in the scripted scene. For example, in a beach scene, the protagonist will train alongside the user, performing similar movements as the user trains, accompanying the user.

[0061] b. The protagonist does not appear: the user does not customize the virtual protagonist image.

[0062] 2. Temperature and humidity control during script training: The default setting here is to always maintain a comfortable state for the human body (moderate moderation, 26°C ± 1°C). However, some users' skin is too sensitive to changes in humidity and temperature. The AI ​​hub can help users automatically and slowly increase or decrease the temperature and humidity during training, and implement exposure therapy for users to reduce sensitivity.

[0063] 3. Oral muscle training tutorial: After the scene is generated, the AI ​​center will automatically generate the corresponding oral muscle training tutorial. This tutorial is divided into 6 dimensions for training:

[0064] a. Lips: Focus on observing the movement of the lips, the location of force, and the response of the lips to pressure during user training.

[0065] b. Tongue position and force: Focus on observing the position of the tongue and its response to pressure.

[0066] c. Gum muscles: Observe the reactions of the gum muscles during biting movements and when touched by external forces.

[0067] d. Cheek muscles: Focus on observing the reaction of the cheek muscles after applying pressure and actively initiating movement.

[0068] e. Jaw: Observe the jaw muscle response to external pressure and active movement, and the position of the jaw when fully relaxed.

[0069] f. Soft palate and hard palate: Focus on observing the palate's response to external pressure and active movement.

[0070] g. Focusing on the above six dimensions, the oral muscle training tutorial will design a variety of training movements for training (the training movements are divided into speech pronunciation, spitting, swallowing, biting, tongue movement, overall oral movement, and subtle movements). For example, a dolphin appears in the ocean and sprays water. The user also needs to imitate the spraying movement and spray the water in the mouth to the target location (a red circle on the screen in front of the user).

[0071] h. The liquids used for training are divided into three types based on viscosity (pure water, aqueous solution with an appropriate amount of starch, and aqueous solution with a large amount of starch). Different viscosities will directly affect the distance and amount of liquid that the user can spray from their mouth, as well as the user's flexibility in performing other movements.

[0072] i. When the liquid sprayed by the user touches the screen, the AI ​​hub will record the position, amount, and the change in the refractive index of the light caused (this is achieved by the AI ​​hub combined with the visual capture algorithm DeepLabCut).

[0073] j. Inner layer (such as Figure 1 ) has a camera that captures video and audio data while the user is training.

[0074] 4. Music Matching: The AI ​​hub matches music to the scripted scene and extracts the music's rhythm, presenting it with specific drum beats or other syllables to assist children in performing specific training based on the rhythm. For example, a beach scene would typically be paired with soothing light music, and the AI ​​hub would present this music's rhythm with the sound of a chime. The user would then swallow or spray the liquid in their mouth to the target location according to the rhythm.

[0075] 5. Determine the confidence incentive mechanism: When the user initially selects "Confidence Mode," the AI ​​Hub will select a specific incentive mechanism based on the scenario when generating a script. For example, in a beach scenario, a small wave will appear after the user sprays liquid at the target location, while in a desert scenario, the small wave will turn into a puff of dust. Different scenarios have different confidence incentive mechanisms. Furthermore, in Confidence Mode, the AI ​​Hub will automatically adjust the training difficulty and proactively adapt to the user's performance to ensure that the user does not feel frustrated.

[0076] 6. Determine the Challenge Evaluation Mechanism: After the user initially selects "Challenge Mode," the AI ​​Center will generate an evaluation report based on the six dimensions of the oral muscle training tutorial. The report includes the following metrics: a. The number of times each dimension was trained, the degree of achievement each time, and the overall achievement rate; b. The user's speech production throughout the training process; and c. The final achievement of the initially set challenge goal.

[0077] Import a pre-designed standard training script (for a detailed description of the script, see "Self-designed training script"). The standard training script is the same as the self-designed training script and contains the following parts:

[0078] 1. Select training mode

[0079] 2. Make requests to the script in 4 ways and then generate the script. The script contains the following elements

[0080] a. Scene Generation

[0081] b. Temperature and humidity control

[0082] c. Oral muscle training tutorial

[0083] d. Music matching

[0084] e. Identify confidence encouragement mechanisms or challenge assessment mechanisms.

[0085] Evaluate users and obtain evaluation results, specifically:

[0086] The evaluation results are divided into two types according to the mode (confidence mode or challenge mode) selected by the user during the script generation phase:

[0087] 1. Confidence Mode Evaluation: This focuses on highlighting the user's progress, growth, and strengths. There is no objective scoring. The evaluation specifically includes the following aspects:

[0088] a. Present audio and video clips of users demonstrating progress during training.

[0089] b. The completion status of the training goals set by the user is displayed. If the goals are not completed, an encouraging animation is displayed to acknowledge the user's progress and encourage them to continue trying and training.

[0090] c. A system that encourages debate: Specific audio and video clips are extracted from the user's training process, and then questions are posed to the user. The questioning format and content may not be optimal, but the user is encouraged to debate. This helps train their speech and improve their oral muscles. For example, the audio and video clips show that the user sprayed water, hitting two red circles, but the question posed to the user is, "Kid, how many red circles were hit? I think it's one, what do you think?" Users are prone to arguing with obvious factual errors. The system guides the user step by step, constantly affirming their arguments, and ultimately both parties reach an agreement, reaffirming and encouraging the user's confidence.

[0091] 2. Evaluation in Challenge Mode: The AI ​​Hub will output an evaluation report after training is complete. The relevant indicators are divided into:

[0092] a. The number of times each of the six dimensions was trained, the degree of achievement each time, and the overall achievement rate.

[0093] b. The user's speech during the entire training process.

[0094] c. The final completion status of the challenge goals initially set.

[0095] d. Debate Encouragement System: This system is consistent with the "Confidence Mode", but after the debate is over, the user's performance will affect the score of the above indicators (improve the above score).

[0096] Figure 3 Describes the artificial intelligence-related technologies that this system mainly relies on.

[0097] Data processing center with ImageBind large model as the core: With ImageBind large model as the processing core, it is responsible for the overall data collection, processing, and summary of this system. The data of information, evaluation, companionship, display and other functions are fully responsible by this data center.

[0098] Virtual image generation technology with Shap-E+OpenAvatar as the core: When generating scripts, this system allows users to customize the generation of a "virtual protagonist image" and relies on the virtual image generation technology with Shap-E+OpenAvatar as the core to achieve these functions.

[0099] DeepLabCut algorithm toolkit: Through artificial intelligence visual capture, this toolkit is mainly responsible for analyzing the user's movements during training and providing data basis for subsequent evaluation.

[0100] The above embodiments are intended to illustrate the present invention, not to limit the present invention. Any solution that is a simple transformation of the present invention falls within the protection scope of the present invention.

Claims

1. An oral muscle training system, characterized by: The invention comprises an oral muscle training space and an oral muscle training method on a computer built in the oral muscle training space, wherein a video module (1) is installed on the top of an inner layer (4) of the oral muscle training space, a temperature and humidity control module (2) is installed on the lower front and rear sides of the inner layer (4) of the oral muscle training space, an outer layer (3) of the oral muscle training space is a solid wall, and the inner layer (4) of the oral muscle training space is composed of 5 LED display screens, and the video module (1), the temperature and humidity control module (2), and the LED display screen are all connected to the computer by wired or wireless communication; The inner layer (4) includes five LED display screens, and the four walls and the top of the inner layer (4) are all LED display screens. There is a gap between the inner layer (4) and the outer layer (3), and the gap is between 50-100 cm. It also includes a debate encouragement system, which is used to extract specific segments of audio and video during user training and ask questions to encourage users to debate. The oral muscle training method comprises the following steps: Step 1: Design your own training script or import a standard training script. The specific steps of the self-designed training script are: Step z11: Select training mode and training goal: the training mode includes confidence mode and challenge mode; Step z12: Input training script generation requirements and generate a training script: The user inputs the training script generation requirements, and the AI ​​hub in the computer generates a training script based on the training script generation requirements. The requirements include text descriptions, pictures, audio, and video. The training script includes elements such as scene generation, temperature and humidity control during script training, oral muscle training tutorials, music matching, and determining a confidence encouragement mechanism or a challenge assessment mechanism. Step 2: Complete training according to the training script; Step 3: Evaluate the user and obtain the evaluation results.

2. The oral muscle training system according to claim 1, wherein: The computer has a data processing hub with the ImageBind large model as its core.

3. The oral muscle training system according to claim 1, wherein: The video module (1) includes an ultra-short-throw projector and a camera. The ultra-short-throw projector is used to make the ground of the inner layer (4) become a screen, and the camera is used to collect the video and audio of the user during the training process.

4. The oral muscle training system according to claim 1, wherein: The temperature and humidity control module (2) is used to control temperature and humidity, and to perform heating or cooling, humidification or drying.

5. The oral muscle training system according to claim 1, wherein: The ground of the inner layer (4) is provided with a drainage device.

6. The oral muscle training system according to claim 1, wherein: The specific import standard training script of step 1 is: Step d11: Select a training mode: the training mode includes a confidence mode and a challenge mode; Step d12: Output the pre-imported standard training script: The standard training script includes the following elements: scene generation, temperature and humidity control during script training, oral muscle training tutorial, music matching, and determination of confidence encouragement mechanism or challenge assessment mechanism.

Images