A vehicle display image generation method, device, medium and product

Abstract

This application provides a method, device, medium, and product for generating vehicle display graphics and text, relating to the field of vehicle interaction technology. The method includes: acquiring inference data related to the emotions of a target user associated with the vehicle, and obtaining user emotional information based on the inference data; generating graphic and text information matching the user's emotional information. This application actively collects multi-source real-time data to perceive user emotions and automatically generates graphic and text information matching the current emotion. It can automatically adapt the graphic and text style to different driving scenarios (such as commuting anxiety, long-distance fatigue, and travel enjoyment), helping to regulate user emotions, alleviate negative driving psychology, and improve driving comfort and interactive experience.

Description

Technical Field

[0001] This application relates to the field of vehicle interaction technology, and more specifically, to a method, device, medium, and product for generating vehicle display graphics and text. Background Technology

[0002] With the continuous improvement of automotive intelligence, in-vehicle infotainment systems have evolved from traditional audio-visual entertainment devices into intelligent interactive hubs integrating navigation, social interaction, office work, and entertainment. Users' demands for personalized and emotional experiences in in-vehicle interfaces are increasing. Currently, the generation and display of text and image information in in-vehicle display systems are relatively simple, such as using static images, simple animated images, and fixed text and images manually downloaded or synchronized by the user. These solutions typically only provide basic visual decoration or information presentation functions, with the text and image content either pre-selected by the user or fixedly provided by the system.

[0003] In related technologies, the vehicle display graphics and text information generated by the above-mentioned existing solutions are in a passive display state. The display content is pre-selected by the user or fixedly provided by the system, and the generation basis comes from preset options. This results in a monotonous in-vehicle display interface experience, a lack of emotional resonance, and difficulty in meeting users' expectations for a smart cockpit experience. Summary of the Invention

[0004] The problem addressed in this application is how to achieve proactive perception and dynamic adaptation of vehicle display graphics and text information to user emotions.

[0005] To address the aforementioned issues, this application provides a method, device, medium, and product for generating vehicle display graphics and text.

[0006] Firstly, this application provides a method for generating vehicle display images and text, comprising: Obtain the emotion-related data of the target user associated with the vehicle, and obtain the user's emotional information based on the data to be inferred; Generate image and text information that matches the user's emotional information.

[0007] Optionally, the data to be inferred includes data collected by at least two sensors.

[0008] Optionally, the data to be inferred includes at least one type of user sentiment data selected from the vehicle's sensor data, the user-side ecosystem data corresponding to the vehicle, and the vehicle's external environment data; and / or, The graphic information includes at least one of the following: images, videos, text, live wallpapers, virtual car model renderings, and cockpit lighting-linked visual scenes.

[0009] Optionally, the data to be inferred includes at least two user emotion data points, and obtaining user emotion information based on the data to be inferred includes: The at least two user sentiment data are fused to obtain the fused data to be inferred; The user's emotional information is obtained by performing reasoning actions based on the data to be fused using a preset emotion classification model.

[0010] Optionally, obtaining the user's emotional information by performing inference actions based on the data to be fused using a preset emotion classification model includes: Using the preset sentiment classification model, based on the task instructions in the preset prompt words, the data to be fused is semantically understood and sentiment analyzed, and the user sentiment information is output according to the output format in the preset prompt words.

[0011] Optionally, generating the image and text information matching the user's emotional information includes: The rendering engine's material parameters are adjusted using a set of material parameters corresponding to the user's emotional information, and the adjusted rendering engine is used to generate graphic and textual information.

[0012] Optionally, the step of adjusting the material parameters of the rendering engine using a material parameter set corresponding to the user's emotional information, and generating graphic information using the adjusted rendering engine, includes: Based on the mapping relationship between each emotion category in the preset emotion category set and the preset material parameters in the rendering engine, the set of material parameters corresponding to the user's emotion information is determined. The material parameters of the rendering engine are adjusted using the material parameter set to obtain the adjusted rendering engine. The adjusted rendering engine performs real-time rendering based on a preset scene model to generate graphic and textual information that matches the user's emotional information.

[0013] Optionally, after generating the image and text information matching the user's emotional information, the method further includes: The graphic information is output through at least one display unit of the vehicle.

[0014] Secondly, the vehicle display graphic generation device of this application includes: The emotion perception and processing unit is used to acquire emotion-related data to be inferred from the target user associated with the vehicle, and to obtain user emotional information based on the data to be inferred. The generation unit generates image and text information that matches the user's emotional information.

[0015] Thirdly, an electronic device according to this application includes a memory and a processor; The memory is used to store computer programs; The processor is used to implement the above-described method for generating vehicle display images and text when executing the computer program.

[0016] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when read and run by a processor, implements the aforementioned method for generating vehicle display images and text.

[0017] Fifthly, this application provides a computer program product, comprising a computer program that, when executed by a processor, implements the aforementioned method for generating vehicle display images and text.

[0018] This application discloses a method, device, medium, and product for generating vehicle display images and text. By acquiring data related to user emotions and inferring user emotional information from this data, it generates matching image and text information. Compared to existing static images or fixed dynamic image solutions, this application no longer relies on passive user selection or system presets. Instead, it actively collects multi-dimensional, multi-source real-time data, enabling the vehicle system to perceive the user's current emotional state and generate emotionally resonant image and text information, such as wallpaper content, based on this emotion. This application can automatically generate image and text information matching the user's true emotions for different driving scenarios (such as commuting anxiety, long-distance fatigue, and travel enjoyment). For example, when the user is identified as anxious or fatigued, soothing and relaxing image and text information can be generated; when a pleasant state is detected, bright and interesting image and text information can be generated. This emotional and personalized visual feedback helps alleviate negative driving emotions, regulate psychological state, and improve driving comfort and safety. In addition, this application automatically completes the entire process of emotion data collection, emotion category reasoning, and graphic information generation without requiring manual settings or selection by the driver or passengers. It also reduces the interactive burden during driving to a certain extent and lowers the risk of distraction caused by operating the vehicle's infotainment interface. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the vehicle display image and text generation method according to an embodiment of this application; Figure 2 This is a schematic diagram of the vehicle display graphic generation device according to an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of this application more apparent and understandable, specific embodiments of this application are described in detail below with reference to the accompanying drawings. Although some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the accompanying drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application.

[0021] It should be understood that the steps described in the method embodiments of this application may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this application is not limited in this respect.

[0022] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc., mentioned in this application are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0023] It should be noted that the terms "a" and "a plurality of" used in this application are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0024] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties. The collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0025] like Figure 1 As shown in the figure, this application embodiment provides a method for generating vehicle display images and text, including: Obtain the emotion-related data of the target user associated with the vehicle, and obtain the user's emotional information based on the data to be inferred; Specifically, real-time data from various functional modules of the vehicle's infotainment system, generated from different sources and related to user emotions (multi-source data), is collected as inference data related to user emotions. This inference data may include, but is not limited to, at least one type of user emotion data. User emotion data may include, but is not limited to, data collected by at least one functional module: the vehicle dynamics module, the environmental perception module, and the voice interaction module. Specifically, the vehicle dynamics module may collect operational data such as vehicle speed and acceleration as user emotion data; for example, repeated rapid acceleration may indicate that the user is in a state of excitement or impatience. The environmental perception module may collect, but is not limited to, at least one type of environmental data such as time, weather, and cabin temperature as user emotion data; for example, high temperatures and traffic congestion can easily induce irritability. The voice interaction module may collect, but is not limited to, at least one type of interactive data, such as the user's voice or the text of user instructions, as user emotion data; for example, a user saying "This road condition is really a headache" accompanied by high speed and rising intonation clearly indicates a negative emotion. Then, the collected at least one type of user emotion data is used as the aforementioned inference data.

[0026] In this embodiment, the target user refers to a natural person associated with the current vehicle. Their specific identity may include, but is not limited to: the vehicle owner (i.e., the registered or long-term user of the vehicle), the driver currently driving the vehicle, the passenger in the front passenger seat, the owner's relatives (e.g., spouse, children, parents, or friends), and other individuals identified as having a temporary or long-term association with the vehicle through vehicle authorization systems such as Bluetooth keys, account logins, and seat memory. The system can identify the current target user through methods such as facial recognition from the in-vehicle camera, seat pressure detection, mobile phone Bluetooth binding information, or an account actively selected by the user. It is worth noting that this embodiment may not require user identification; as long as data related to the user's emotions can be obtained, subsequent processing can proceed. In multi-passenger scenarios, data from the primary driver can be collected first, or the user requiring service can be confirmed through voice interaction.

[0027] The system collects data related to user emotions from various sources (i.e., multi-source data) generated in real time by different functional modules of the vehicle's infotainment system as inference data. This inference data can include at least one type of user emotion data. User emotion data refers to any data information that can directly or indirectly reflect, characterize, or infer the user's current emotional state. This data can be in the form of structured numerical values ​​(such as sensor readings) or unstructured text or audio features. Specifically, user emotion data can originate from, but is not limited to, the following functional modules: vehicle dynamics module, environmental perception module, and voice interaction module.

[0028] For example, two user sentiment data from the vehicle dynamics module: the average vehicle speed in the past minute was 15 km / h, and the number of sudden braking incidents in the same period was 5. Combining these two data points, it can be inferred that the user is in a stop-and-go traffic jam environment and may be in a state of anxiety.

[0029] In another example, user emotion data from the environmental awareness module: the current time is 19:30 (evening), the weather is cloudy, and the cabin temperature is 21 degrees Celsius. These data, taken individually, do not directly point to a certain emotion, but as context, they can help with inference. For example, if it is evening and the temperature is comfortable, and there is no other negative data, the user may be relatively relaxed.

[0030] In another example, user sentiment data from the voice interaction module: a user says, "I'm in a good mood today, going for a drive after work," and the semantic sentiment of this text is positive. The system can use one or more of the above user sentiment data as data to be inferred.

[0031] As one example, when the data to be inferred includes at least two user emotion data, the at least two user emotion data can be fused according to a preset data structure (such as, but not limited to, JSON data structure) to obtain the data to be inferred; specifically, the data to be inferred can be constructed based on at least two user emotion data into a unified data format that includes module name, data type, data content and unit.

[0032] Generate image and text information that matches the user's emotional information.

[0033] Specifically, based on the obtained user emotional information, the corresponding graphic and text content is determined so that the final displayed graphic and text content is related to the user's current emotional state in terms of style, color, theme, or emotional tone.

[0034] In a preferred embodiment, after obtaining user emotional information, such as emotional classification tags and emotional dimension values, a correspondence is established between the emotional information and image / text features, which are then transformed into visualized image / text output. Image / text features may include the dominant color tone, saturation, contrast, pattern elements, and dynamic effects of the image, as well as the semantic emotion and tone style of the text. In practical application generation, different technical solutions can be adopted according to different system architectures and performance requirements, such as pre-set material matching, parametric dynamic generation, or real-time creation based on models.

[0035] In one implementation, a matching method based on a pre-built material library can be used. Specifically, an emotion-image-text mapping library is pre-built, containing various emotion categories such as happiness, calmness, sadness, and excitement, each with corresponding images, animated images, text, or combinations of these elements. When the target user's emotion information is obtained, this emotion information is matched against the emotion categories in the library to retrieve the corresponding image-text materials. These materials are then selected, overlaid, or spliced ​​to form the final image-text information. For example, if the user's emotion information is "happy," a brightly colored image containing sunlight elements and the text "Feeling happy, have a safe trip" are matched; if the user's emotion information is "calm," a soft-toned, scenic image and a soothing sentence are matched. This method is simple to implement, responds quickly, and is suitable for scenarios with high real-time requirements but limited computing resources.

[0036] In another implementation, a dynamic synthesis method based on parameterized rules can be used. Specifically, a mapping function between emotional information and visual attributes of text and images is predefined. These visual attributes include, but are not limited to, dominant hue (hue value), saturation, brightness, graphic shape parameters, particle motion speed, and text emotional tendency values. After obtaining the user's emotional information, it is substituted into the mapping function as input parameters to calculate specific attribute values ​​in real time. Then, the basic drawing or animation module is called to dynamically synthesize the text and image content based on these attribute values. For example, emotional information can be mapped to two continuous dimensions: valence and arousal. Valence determines hue; for example, positive emotions lean towards warm colors, and negative emotions lean towards cool colors. Arousal determines saturation and the rate of dynamic change; for example, high arousal corresponds to high saturation, rapid flashing, or motion, while low arousal corresponds to low saturation and slow changes. This method can generate continuously changing and unique text and image information, avoiding the finiteness and discreteness of a pre-built material library.

[0037] In another implementation, a real-time creation approach based on generative models can be adopted. Specifically, pre-trained multimodal generative models, such as large language models, image generation models, or a combination of both, can be used. User sentiment information is taken as conditional input, and the model directly generates matching images and text. Specifically, sentiment tags or sentiment vectors can be input into the model as part of the prompt words. The model outputs corresponding image data and natural language descriptions, which are then post-processed, such as size cropping, color correction, and text layout, to form the final image and text information. This method offers the highest degree of personalization and can produce an infinite variety of image and text results, suitable for users with high demands for personalized experiences.

[0038] In another implementation, an optimization based on the pre-set material library matching method can be achieved by combining templates and parameter adjustments. Instead of using complete image and text materials as the matching unit, templates and adjustable parameters are stored separately. Each template defines the basic layout and theme elements of the image and text, while reserving several adjustable parameters, such as background color, element color, and animation speed. Emotional information is used not only to select templates but also to determine the values ​​of these adjustable parameters. After obtaining the user's emotional information, the system first selects a basic template based on the emotional information, then calculates the adjustment values ​​of various parameters in the template based on the emotional information, and finally applies the parameter adjustments to the template to generate the final image and text information. This method combines the speed advantage of the matching method with the continuous adjustment capability of the parameterized method, improving the precision of the matching while ensuring response efficiency.

[0039] In another implementation, a dynamic selection method based on ranking and recommendation can be used. Specifically, this method can be adopted when the system has a large amount of image and text materials, such as an online wallpaper library or a user's personal image library, and it is not possible to pre-label each material with emotional tags. The visual feature vector and textual semantic vector of each image and text material are pre-extracted, while the user's emotional information is converted into a query vector. Through vector similarity calculation, the system retrieves several candidate materials from the material library that have the highest similarity to the current emotional information in real time. These are then ranked in conjunction with the user's historical preferences, and finally, the image and text information with the highest ranking is selected for display.

[0040] In this application, the above-mentioned implementation methods can be used individually or in combination depending on factors such as the system's computing power, real-time requirements, and the richness of material resources, and will not be elaborated further here.

[0041] In one embodiment, the multi-source data of each module in the vehicle infotainment system is encapsulated as a `data_source` array structure: each element in the array corresponds to a data module, containing a `module_name` field to identify the module name, such as "Vehicle Status", "Time", or "Voice Module"; the `datas` array under each module stores multiple data items, each containing three fields: `type` (data type), `content` (data content), and `unit` (data unit). For example, the "Vehicle Status" module contains: data type "vehicle speed", content "60", unit "km / h"; data type "acceleration", content "50", unit "60"; and data type "acceleration", content "50", unit "60". The "Time" module contains: data type "Time", content "18:00", unit left blank; data type "Weather", content "Cloudy", unit left blank; data type "Cabin Temperature", content "21", unit "degrees Celsius". The "Voice" module contains: data type "Dialogue Text", content "The weather is so nice today, the night is so beautiful", unit left blank. After serializing the above structured data into a buffer suitable for transmission, it can be sent to the preset model through the interface mcp_client_sendToAIModel(buffer); after the preset model completes processing, it returns the standardized inference result. For example, the format of the standardized inference result can be {"Inference_result": {"value":} The value field of the string "0" can take values ​​of 0, 1, 2, and 3, corresponding to the four emotion categories of "joy," "anger," "sorrow," and "happiness," respectively. It should also be noted that in actual implementation, the number of emotion categories and specific labels can be flexibly defined according to the system design, such as "calm," "excitement," "anxiety," and "romance," or multi-dimensional emotion vectors (such as effectiveness value and arousal level) can be used to replace discrete labels. Any output format that can represent the user's emotional state and be recognized by the downstream rendering engine and mapped to material parameters falls within the protection scope of this application.

[0042] The vehicle display image and text generation method of this embodiment acquires data related to user emotions for inference, and infers user emotional information based on this data, thereby generating matching image and text information. Compared with existing static images or fixed dynamic image solutions, this application no longer relies on the user's passive selection or system preset, but actively collects multi-dimensional and multi-source real-time data, enabling the vehicle system to perceive the user's current emotional state and generate image and text information with emotional resonance based on this emotion, such as wallpaper content. This application can automatically generate image and text information that matches the user's true emotions for different driving situations (such as commuting anxiety, long-distance fatigue, and travel pleasure). For example, when the user is identified as being anxious or fatigued, soothing and relaxing image and text information can be generated; when a pleasant state is identified, bright and interesting image and text information can be generated. This emotional and personalized visual feedback helps alleviate the user's negative driving emotions, regulate their psychological state, and improve driving comfort and driving safety. In addition, this application automatically completes the entire process of emotion data collection, emotion category reasoning, and graphic information generation without requiring manual settings or selection by the driver or passengers. It also reduces the interactive burden during driving to a certain extent and lowers the risk of distraction caused by operating the vehicle's infotainment interface.

[0043] Optionally, the data to be inferred includes data collected by at least two sensors.

[0044] In this optional embodiment, information related to user emotions is captured from different dimensions using data from multiple sensors, thereby improving the accuracy and robustness of emotion inference. Specifically, various sensors are deployed inside and around the vehicle, such as in-vehicle cameras for capturing driver facial expressions, microphones for capturing voice tone, biosensors for monitoring heart rate or skin conductance responses, such as sensing electrodes on the steering wheel or seat, vehicle speed sensors, acceleration sensors, and steering angle sensors for collecting driving behavior data, and temperature sensors, light sensors, and rain sensors for sensing the external environment. Each of these sensors outputs data reflecting the user's or environment's state. However, data from a single sensor often has limitations; for example, cameras may fail in low light, voice input is ineffective when the user is not speaking, and biosensors may be affected by the way they are worn. Therefore, this embodiment uses data from at least two sensors as the data to be inferred, leveraging the complementarity of multi-source data to improve the reliability of emotion judgment. At the data processing level, the data collected by at least two sensors can be of the same modality, such as audio collected by microphones in two different locations, or of different modalities, such as image plus speech, speech plus biosignals, driving behavior plus ambient light, etc. After synchronizing and structuring these multi-source data, they are fed into the subsequent sentiment inference module. By employing data from at least two sensors, the technical solution of this application can maintain effective sentiment perception capabilities even when a single sensor fails or the data quality is poor. It also provides a richer information foundation for subsequent fusion inference, thereby improving the robustness and accuracy of the overall solution.

[0045] For example, voice interaction content captured by the in-vehicle microphone, such as "The traffic is really frustrating today," and changes in grip force detected by the pressure sensor on the steering wheel can be collected simultaneously. When the voice content expresses negative emotions and the grip force increases significantly, both indicate that the user is in a state of anxiety or anger, and the system generates more accurate emotional information accordingly. Another example is the combination of smile intensity captured by the facial camera and changes in vehicle speed sensor data: if the user smiles while the vehicle speed is stable, it is tended to be judged as a happy state; if the smile is accompanied by rapid acceleration or deceleration, it may indicate excitement or tension.

[0046] Optionally, the data to be inferred includes at least one type of user sentiment data selected from the vehicle's sensor data, the user-side ecosystem data corresponding to the vehicle, and the vehicle's external environment data; and / or, The graphic information includes at least one of the following: images, videos, text, live wallpapers, virtual car model renderings, and cockpit lighting-linked visual scenes.

[0047] Specifically, in this embodiment, the data to be inferred covers at least one or more of the following data sources: The first category is the vehicle's own sensor data, which consists of dynamic operating parameters collected in real time by the vehicle's various electronic control units and onboard sensors. Specifically, the vehicle's own sensor data includes, but is not limited to: real-time vehicle speed data collected by the vehicle speed sensor, longitudinal and lateral acceleration recorded by the acceleration sensor, wheel speed measured by the wheel speed sensor and the vehicle attitude calculated from it, steering wheel angle and angular velocity monitored by the steering angle sensor, brake master cylinder pressure obtained by the brake pressure sensor, accelerator pedal position sensor outputting the opening value, tire pressure and temperature provided by the tire pressure monitoring system, engine or motor control system feedback on speed and torque output, and vehicle yaw rate provided by the inertial measurement unit. This data exists in structured numerical form and can directly reflect the vehicle's driving status and driving behavior characteristics, thereby indirectly inferring the user's emotional tendencies. For example, frequent rapid acceleration and braking may correspond to anxious emotions, while smooth accelerator and brake often correspond to a relaxed state.

[0048] The second category is the user-end ecosystem data corresponding to the vehicle, which is the personalized data of the user obtained through user terminal devices bound to the vehicle, such as smartphones, smartwatches or cloud account services, such as the user's schedule, music preference settings, health monitoring data, etc. This data is independent of the vehicle's sensors and can provide information related to emotions from the dimensions of the user's daily behavior and life status.

[0049] The third category is external environmental data of the vehicle, which is real-time environmental status information obtained through onboard environmental perception sensors, such as rain sensors, light sensors, temperature sensors, or networked environmental data services. This includes weather conditions, such as sunny, rainy, snowy, foggy, etc., ambient temperature, geographical location, such as whether it is in a congested city center, scenic area, highway, and road conditions, such as smooth traffic, slow traffic, congestion, etc. Although the external environment does not directly reflect the user's emotions, it is often an important trigger for emotional changes and can be used as contextual information to assist in emotional reasoning.

[0050] The fourth category is the real-time dialogue text obtained through the in-vehicle voice interaction module. This refers to the text information obtained after the voice signals generated during the voice dialogue between the in-vehicle occupants and the vehicle's infotainment system are recognized and converted. As unstructured semantic data reflecting the user's real-time tone, this dialogue text directly contains the user's current emotional tendency and sentiment. For example, a user's exclamation, "The weather is so nice today, it makes me feel so much better," clearly points to a positive emotion.

[0051] The aforementioned multi-source sensor data, from the dimensions of vehicle operation, user behavior, external environment, and voice interaction, jointly construct a relatively complete description of the driving scenario. This data will be sent as a whole into the subsequent fusion processing flow as the data to be inferred, so as to ensure that the emotion classification model can make more accurate inferences and judgments about the user's current emotional state based on multi-dimensional information.

[0052] In this optional embodiment, by unifying multi-dimensional information such as vehicle sensor data, user-end ecosystem data, external environment data, and in-vehicle voice dialogue content as the data source to be inferred, the emotion classification model can comprehensively consider different levels of emotion-inducing factors such as vehicle operating status, user behavior preferences, external environmental context, and voice interaction semantics, thereby effectively improving the comprehensiveness and accuracy of inferring the user's current emotional state and providing a more reliable data foundation for the personalized adaptation of mood wallpapers.

[0053] In another preferred embodiment of this application, the vehicle system includes multiple functionally independent data modules. Each data module is responsible for collecting sensor data of different dimensions. For example, the vehicle dynamics module collects vehicle speed and acceleration data, the environmental perception module collects time, weather, and cabin temperature data, and the voice interaction module collects dialogue text data. Each data module independently generates multi-source sensor data with timestamps according to its own collection cycle or event triggering mechanism, and reports it to the data receiving service on the vehicle terminal in real time. The data receiving service maintains a global data buffer, which is used to temporarily store the latest sensor data from each data module at the current moment, and uses all the data in the buffer as the data to be inferred. Since the data reporting time of each data module is independent, the data fields in the buffer always remain the most recently reported valid values ​​of each module. When any data module generates and reports new sensor data, the data receiving service updates the corresponding data field in the buffer for that module, thereby dynamically refreshing the content of the data to be inferred. Each update to the data buffer is configured to trigger a re-execution of the subsequent sentiment inference process. This means the system sends the updated data to be inferred to the fusion processing stage and initiates a new step of sentiment category matching using a preset sentiment classification model. This ensures that the sentiment inference results can respond promptly to dynamic changes in user status or driving context. During the data update process, if a data module fails to report valid data on time due to sensor malfunction, communication interruption, or signal anomaly, the data receiving service maintains the previously successfully received valid value for the corresponding data field of that module, without performing null value filling or data clearing operations. Subsequent sentiment inference processes are executed only based on the valid data already present in the current buffer. Missing or abnormal data modules do not affect the continuation of the overall inference process, thus ensuring the system's robustness and availability even when some data sources fail. In this embodiment, the vehicle-mounted data fusion processing mechanism adopts a dynamic incremental update strategy with module names as index keys. Specifically, the global data buffer maintained by the data receiving service uses the module_name field of each object in the data_source.modules list as a unique identifier. Each data module reports data according to its own independent collection cycle or event triggering sequence. The system locates the data area of ​​the corresponding module in the buffer based on the module_name to which the reported data belongs, and incrementally overwrites or appends the corresponding data items in a way similar to hash table key-value updates. The timing of each module is uncoupled and they can be accessed asynchronously and freely. When a data module loses data or times out due to sensor abnormality or communication failure, the system does not perform any update operation on the data area corresponding to that module in the buffer, that is, it maintains its original data state unchanged. The subsequent sentiment inference process is only executed based on the module data that already exists in the current buffer.

[0054] Optionally, the data to be inferred includes at least two user emotion data points, and obtaining user emotion information based on the data to be inferred includes: The at least two user sentiment data are fused to obtain the fused data to be inferred; The user's emotional information is obtained by performing reasoning actions based on the data to be fused using a preset emotion classification model.

[0055] Specifically, user emotion data can come from various data sources mentioned above, such as facial expression feature values ​​collected by an in-vehicle camera, voice tone parameters collected by a microphone, grip strength or posture change values ​​output by a pressure sensor on the steering wheel or seat, driving behavior data such as vehicle speed and acceleration, the degree of stress in the user's schedule or music preference type in the user's ecosystem data, and weather and road conditions in the external environment data. In one embodiment, the two user emotion data can be exactly the same type or different types; they can come from the same sensor, such as two frames of facial expression feature values ​​collected continuously by the same camera, or two sentences of voice text collected by the same microphone at different times, or they can come from multiple different sensors, such as facial expression features collected by an in-vehicle camera and grip strength values ​​collected by a steering wheel pressure sensor, or the number of rapid accelerations measured by a vehicle speed sensor and complaining text recognized by a voice module; they can be structured numerical data, such as heart rate values ​​and acceleration fluctuation variance, or unstructured semantic data, such as dialogue text and voice tone parameters; or two different feature parameters provided by the same data source, such as the angle of the corners of the mouth and the degree of eyebrow curvature extracted simultaneously from a facial image. This condition is met as long as there are two independent data units that can be used for emotion inference. After obtaining at least two user emotion data points, user emotion information will be obtained through subsequent fusion processing and emotion classification model inference. This leverages the information redundancy or complementarity brought by multiple data points to improve the accuracy and robustness of emotion judgment.

[0056] This embodiment sets up a data fusion processing mechanism on the vehicle-mounted system. During operation, the vehicle-mounted system collects inference data from multiple functional domains in real time, including but not limited to vehicle dynamic operation data, environmental status perception data, and in-vehicle interaction behavior data. However, the information output from these data sources has different formats and semantic types, including both numerical parameters and unstructured text descriptions. The data fusion processing mechanism, based on a pre-agreed unified data description framework, classifies, integrates, and standardizes the original data from different sources. It merges the originally scattered and heterogeneous multi-source inference data into a unified and semantically complete inference fusion data package, which is then temporarily stored in a local buffer. When the data to be fused in the data fusion buffer is updated, the vehicle system sends the complete data packet to be fused via a pre-defined cloud communication link to a pre-defined emotion classification model deployed in the cloud. Upon receiving the data packet, the cloud-based emotion classification model executes the inference action based on the system prompts configured during initialization. These system prompts predefine the model's analytical role, analytical objective, rules for parsing the input data structure and content, the range of available emotion categories, and the output format of the inference results. When parsing the data packet, the model extracts structured numerical features and unstructured semantic features reflecting the user's driving state and interaction tone, and combines these features for matching and discrimination within a pre-defined emotion category set. After inference, the model determines at least one emotion category that best matches the user's emotional state represented by the current data to be fused, and uses this category as the user's emotional information, outputting the inference results in a pre-defined format for subsequent wallpaper rendering or in-vehicle atmosphere control processes.

[0057] In one embodiment, the prompt word is composed of four parts: identity definition, task instruction, constraints, and output requirements. The identity definition part sets the preset sentiment classification model as a senior contextual sentiment analyst; the task instruction part explicitly requires the model to analyze the user's current emotional state based on the user's real-time driving data and the tone of the conversation; the constraints part stipulates that the input data must follow a preset nested JSON data structure, that is, with data_source as the root object, under which the modules array carries the data of each functional module. Each module is identified by the module_name field, and the datas array contains the specific data items under the module. Each object in datas uses the type field to represent the physical meaning of the data, the content field to represent the current value or content, and the unit field to represent the unit of measurement. On the other hand, it limits the output dimension of sentiment analysis to four sentiment categories: joy, anger, sorrow, and happiness; the output requirements part stipulates that the model must return the inference results in a preset JSON format, that is, with Inference_result as the root structure, the value field in which the inference result is stored, and the value of value is limited to integers from zero to three, corresponding to the four sentiment states of joy, anger, sorrow, and happiness, respectively. The aforementioned prompts are sent to the preset emotion classification model via a preset communication interface during the function initialization phase to complete the unified preset configuration of the model's analysis identity, task scope, input / output format, and emotion dimension. The emotion description dimensions mentioned in the constraints are not limited to the four categories of joy, anger, sorrow, and happiness in this embodiment; other prompts used to describe human emotional states should also be understood as falling within the scope of this application's technical solution.

[0058] In another embodiment, the preset sentiment classification model can be a large-scale language model deployed in the cloud. This large-scale language model is a natural language processing model built on a deep neural network and has a large number of parameters. Simultaneously, other intelligent algorithms with text sentiment analysis capabilities can also serve as specific implementations of this sentiment classification model. In this embodiment, the large-scale language model can be configured with preset system prompts. These prompts include the setting of the model's analysis role, a description of the required analysis task, constraints on the structure and format of the input data, limitations on the range of output sentiment dimensions, and explicit requirements for the format of the inference results. Through the loading of these prompts, the large-scale language model is configured as an inference model specifically used to identify the user's emotional state from multi-source input information that integrates vehicle operation data, environmental status data, and in-vehicle interaction data. Upon receiving the fused data reported in real-time from the vehicle's infotainment system, the model performs matching judgments within a preset sentiment set according to the preset analysis role and task instructions, and outputs the matching results in an inference result containing sentiment category identifiers in a pre-defined format.

[0059] In this optional embodiment, by performing unified format fusion processing on multi-source heterogeneous data such as vehicle dynamics, environmental perception, and in-vehicle interaction, scattered information is integrated into semantically complete fusion data to be inferred, effectively improving the comprehensiveness and accuracy of the basis for emotion judgment. On this basis, the fusion data is sent to a cloud-based preset emotion classification model, and the deep semantic understanding capability of a large language model is used to perform inference matching. This can more accurately capture the user's emotional tendencies from limited in-vehicle data, providing reliable decision support for realizing the proactive adaptation of in-vehicle wallpapers to the user's status.

[0060] Optionally, obtaining the user's emotional information by performing inference actions based on the data to be fused using a preset emotion classification model includes: Using the preset sentiment classification model, based on the task instructions in the preset prompt words, the data to be fused is semantically understood and sentiment analyzed, and the user sentiment information is output according to the output format in the preset prompt words.

[0061] Specifically, a pre-defined emotion classification model deployed in the cloud can receive and load system prompts sent from the vehicle's infotainment system during the initialization phase. These prompts contain task instructions that explicitly require the model to analyze the user's current emotions based on real-time driving data and conversational tone. When the vehicle's infotainment system sends the fused data to be inferred to the pre-defined emotion classification model, the model parses the input data according to the loaded task instructions. First, it extracts driving situation features reflected by structured information such as vehicle dynamic values ​​and environmental state parameters from the fused data. Simultaneously, it extracts semantic and emotional features from unstructured content such as voice dialogue text. Then, based on its natural language understanding and contextual reasoning capabilities, it comprehensively analyzes these features to determine the user's emotional tendency in the current driving scenario and interaction context. It then matches these features within a pre-defined set of emotion categories, identifying at least one emotion category that best matches the current fused data as the user's emotional information. After completing the inference, the model generates the inference result according to the output format specified in the system prompts, thus completing the inference output from multi-source fused data to the emotion classification result.

[0062] In this optional embodiment, by introducing system prompts containing clear task instructions into the preset emotion classification model, the large language model can perform semantic understanding and emotion analysis on the data to be reasoned, which integrates driving data and environmental context, and directly return the emotion classification results that can be parsed by the vehicle terminal according to the preset standardized output format. This ensures that the reasoning process from multi-source heterogeneous data to standardized emotion categories has task focus and output consistency, providing a reliable structured input for rapid response of subsequent scene rendering or atmosphere control.

[0063] Optionally, generating the image and text information matching the user's emotional information includes: The rendering engine's material parameters are adjusted using a set of material parameters corresponding to the user's emotional information, and the adjusted rendering engine is used to generate graphic and textual information.

[0064] Specifically, the rendering engine on the vehicle's infotainment system is pre-configured with multiple sets of material parameters. Each set of material parameters is associated with a different preset emotion category. Each set of material parameters defines the visual attribute configuration of each rendered object in the scene under the corresponding emotional atmosphere, including but not limited to color parameters, lighting parameters, texture representation parameters, and dynamic effect parameters.

[0065] In a preferred embodiment, the material parameter set includes, but is not limited to, the following types of parameters: (1) Color parameters, used to control the main color tone, saturation, brightness, and color matching relationship of each object in the rendered image. Specifically, it may include: base color (RGB or HSV value), ambient color, diffuse color, highlight color, color mixing weight, gradient start color and end color, etc. For example, under the emotion category of "joy", the color parameters can be configured as high brightness and high saturation orange-yellow or pink; under the emotion category of "sorrow", it is configured as low saturation, cool or grayscale tone. (2) Lighting parameters, used to define the light source attributes in the scene, affecting the brightness and atmosphere of the image. Specifically, it includes: ambient light intensity and color, directional light angle intensity and color, point light source position intensity radius and attenuation curve, brightness and color of self-illuminating material, softness and transparency of shadow, etc. For example, in the "calm" state, soft diffuse light can be used, the light angle is gentle, and the shadow edge is blurred; in the "excited" state, dynamic flashing point light source or pulsed backlight effect can be added. (3) Texture rendering parameters, used to control the visual texture features of the material surface. Specifically, these include: texture map type, such as brushed metal, carbon fiber, frosted glass, leather, wood grain, etc.; texture scaling ratio; texture tiling method; texture blending degree; intensity of normal map; contrast of roughness map, etc. For example, the "movement" emotion can be matched with high reflective textures of carbon fiber or brushed metal; the "warmth" emotion can be matched with plush or warm wood textures. (4) Dynamic effect parameters, used to define the movement and change patterns of elements in the material or scene. Specifically, these include: emission rate of particle system, particle size, color change curve, particle movement direction and speed; flow speed and direction of material, such as water ripples, cloud drift; pulsation frequency and amplitude of self-illumination intensity; speed curve of rotation or scaling animation; rate of color cycle change, etc. For example, when "joyful", the flowing light particles can be configured to drift at high speed and the light spot flashes at a relatively fast frequency; when "sad", the slowly falling particles or almost still texture flow can be configured. (5) Material surface property parameters, such as: metallicity (adjusts whether the material appears metallic or non-metallic), roughness (affects the range of high light scattering; smooth surfaces have low roughness, rough surfaces have high roughness), transparency (controls the light transmittance of the material), refractive index (determines the degree of refraction of light when passing through the material), self-illumination intensity (makes the material emit light itself without relying on external light sources), and normal perturbation intensity (simulates the surface bumps and dents without changing the geometry). (6) Environmental linkage parameters, i.e., instruction parameters for interacting with peripherals, such as: sending the current material's primary color value to the ambient light via the bus, synchronizing the frequency of dynamic effects to the control parameters of seat vibration or air conditioning fan speed, etc. These parameters are not directly used for rendering, but as an extension of the material parameter set, they enable other actuators in the vehicle to work together.By combining and configuring the above parameters, this embodiment can quickly call the corresponding material parameter set according to the user's emotional information, adjust the output effect of the rendering engine in real time, and generate graphic and text information that highly matches the emotional atmosphere.

[0066] Specifically, the in-vehicle infotainment system pre-builds a basic scene with various adjustable material parameters and visual state configurations. Developers configure corresponding material parameter sets for each preset emotion category based on different emotional atmosphere requirements, establishing a mapping relationship between these parameter sets and the emotion category. For example, material parameter sets representing bright, warm visual effects are bound to the emotion category "joy," while material parameter sets representing gloomy, cold visual effects are bound to "sorrow," and so on. For instance, an emotion inference is performed using a preset emotion classification model. The inference result, containing user emotion information, is sent back to the rendering instance on the in-vehicle infotainment system via a communication channel. Upon receiving the inference result, the system parses it, extracting numerical or symbolic information identifying the user's emotion. Once the system obtains the user's emotion information, it searches for the corresponding material parameter set in the locally pre-built mapping relationship based on the category, and loads the configuration data from that parameter set into the current rendering pipeline of the 3D rendering engine to update the material properties, lighting parameters, and dynamic effect parameters of the basic scene. The rendering engine performs real-time rendering calculations based on the updated parameter set, generating visual content that matches the user's emotional state and visual atmosphere. Furthermore, the generated content can be pushed to the vehicle's central control display as a mood wallpaper, and / or sent via the vehicle's bus to the ambient lighting system to adjust the color temperature, brightness, or dynamic change mode of the ambient lights, ensuring the overall lighting environment inside the vehicle remains consistent with the current emotional category. Thus, through the linkage between cloud inference and preset scene parameters in the vehicle, personalized emotional adaptation to the in-vehicle visual output terminal is achieved.

[0067] Optionally, the step of adjusting the material parameters of the rendering engine using a material parameter set corresponding to the user's emotional information, and generating graphic information using the adjusted rendering engine, includes: Based on the mapping relationship between each emotion category in the preset emotion category set and the preset material parameters in the rendering engine, the set of material parameters corresponding to the user's emotion information is determined. The material parameters of the rendering engine are adjusted using the material parameter set to obtain the adjusted rendering engine. The adjusted rendering engine performs real-time rendering based on a preset scene model to generate graphic and textual information that matches the user's emotional information.

[0068] Specifically, a basic scene model is pre-built in the rendering engine of the vehicle-mounted system. This scene model contains multiple rendering elements whose visual effects can change with parameter variations, such as the surface material of the vehicle body, the color temperature and direction of ambient lighting, and the particle state of the weather system in the background. Simultaneously, multiple sets of material parameter sets are pre-configured in local storage. Each set of material parameter sets is mapped to a preset emotion category. Each material parameter set contains one or more normalized material parameters. In a specific implementation, each material parameter in the set can be expressed in a normalized numerical form. Normalized material parameters refer to mapping the value range of various visual attributes to a standard interval, such as [0, 1], so that parameters of different dimensions can be combined, compared, or interpolated at the same scale, while reducing dependence on specific physical units or the numerical range within the rendering engine. Specifically, the color parameters (such as RGB components), metallicity, roughness, self-illumination intensity, and dynamic particle emission rate contained in the material parameter set, which originally had different physical meanings and value ranges, are normalized. For example, the self-illumination intensity is normalized from 0 to 100. Linear mapping to [0, 1], or directly using metallicity and roughness (which are naturally within [0, 1]), allows all parameters to be expressed within a unified range. Using normalized material parameters facilitates a smooth transition between parameter sets corresponding to different emotion categories. When a user's emotional information gradually changes from "joy" to "calm," the system can perform linear interpolation or Bezier interpolation on the normalized parameters, and then denormalize them back to the values ​​actually needed by the engine, achieving a continuous and natural visual evolution effect and avoiding abrupt jumps during emotion switching. On the other hand, normalized parameters facilitate cross-parameter comparison and fusion. For example, when considering the response of multiple parameters to user emotions, weighted summation or Euclidean distance calculation can be performed directly, without being affected by dimensions.

[0069] In a specific example, the material parameter set can include multiple normalized parameters such as normalized hue (0 for cool, 1 for warm), normalized saturation (0 for grayscale, 1 for full color), normalized roughness (0 for specular, 1 for fully rough), and normalized self-illumination pulsation frequency (0 for still, 1 for fastest flicker). When the user's emotional information is "joy," the hue is set to 0.9, saturation to 0.9, roughness to 0.2, and pulsation frequency to 0.8; when the emotional information changes to "sorrow," the hue is set to 0.2, saturation to 0.2, roughness to 0.7, and pulsation frequency to 0.1. All parameters are expressed within the normalized range, and the underlying driver program then calculates the actual rendering instructions based on the normalized values. In this way, the material parameter set contains one or more normalized material parameters, which simplifies the parameter mapping logic between different emotional categories and provides a unified numerical basis for dynamic and continuous emotional adaptation.

[0070] When the vehicle's infotainment system receives the inference result data packet returned by the preset emotion classification model via a preset communication protocol, and parses out the numerical identifier representing the user's emotional information, it searches for the corresponding material parameter set identifier in a preset mapping table based on this identifier. Subsequently, the vehicle's infotainment system calls the corresponding parameter loading interface to write the configuration data of each item in the material parameter set into the current parameter register of the rendering pipeline, thereby updating the material properties, light source properties, and special effects properties of each rendered object in the basic scene. Based on the updated parameters, the rendering engine performs real-time rendering calculations on the preset scene model, generating visual content that corresponds to the user's emotional information. This visual content can be output to the vehicle's infotainment display as a mood wallpaper, and / or the corresponding color control signal can be sent to in-vehicle environmental devices such as ambient lights via the in-vehicle communication network to adjust their illumination state to coordinate with the current emotional atmosphere, thereby achieving diversified emotional adaptation of the in-vehicle visual environment.

[0071] Optionally, the mood wallpaper includes at least one of 2D mood wallpaper, 3D mood wallpaper, color rendering features of the vehicle's virtual car model, and color linkage features of the cabin lighting.

[0072] Specifically, when the mood wallpaper is presented as a 2D mood wallpaper, the system retrieves the corresponding static or dynamic image from a preset 2D wallpaper image library based on the user's emotional information for display; when the mood wallpaper is presented as a 3D mood wallpaper, the system dynamically adjusts the material parameters, lighting, and animation effects in the 3D scene based on the user's emotional information to generate a corresponding three-dimensional dynamic image; when the mood wallpaper is presented as the color rendering characteristics of a virtual car model, the system changes the body color, metallic texture, or texture effects of the virtual car model based on the user's emotional information, so that the virtual car model presents a visual style that matches the emotion; when the mood wallpaper is presented as the color linkage characteristics of the cabin lighting, the system maps the user's emotional information to the corresponding light color and brightness parameters, and controls the ambient light strips, footwell lights, and other lighting devices in the cabin to change in linkage according to these parameters. In this embodiment, the above-mentioned various mood wallpapers can be used individually or in any combination to achieve multimodal emotional expression.

[0073] In one embodiment, after the vehicle system starts up and loads the 3D wallpaper instance, it first sends the pre-arranged system prompts to the preset sentiment classification model through the MCP Client interface to complete the pre-setting of the sentiment classification task. At the same time, the vehicle system starts a real-time data receiving service to receive the data_source data reported by each module of the vehicle system in real time, and aggregates the data of each module into a total data buffer for fusion according to the preset data structure format. Whenever the data in the buffer is updated, the complete fused data in the current buffer is sent to the preset sentiment classification model through the MCP Client interface for the model to perform subsequent sentiment inference processing.

[0074] In one embodiment, after receiving the fused data sent by the vehicle's infotainment system via the MCP Client interface, the preset emotion classification model performs inference operations and returns the inference results to the 3D instance on the vehicle's infotainment system via the MCP Server interface according to a preset output format. Upon receiving the inference results, the 3D instance parses the Inference_result structure and extracts the value of the value field. Based on the correspondence between the value and the emotion category, it maps the value to the material parameter set bound to the corresponding emotion category (such as joy / anger / sorrow / happiness) within the 3D engine. Then, the updated material is rendered and output to the vehicle's infotainment screen in real time, thereby achieving the technical effect of the vehicle's 3D wallpaper dynamically responding to changes in the user's mood.

[0075] For example, when the vehicle infotainment system 3D instance is launched, the system starts a data receiving service to collect data from various modules in real time. This includes data from the vehicle dynamics module (vehicle speed 60km / h, acceleration 30m / s), the time module (time 19:06, cloudy weather, cabin temperature 21 degrees Celsius), and the voice module (dialogue text "The weather is so nice today, the night is so beautiful"). These data are integrated according to the preset data_source unified JSON data structure and sequentially assigned to the corresponding module_name module under the modules array. Each data item has clearly defined type, content, and unit fields, ultimately completing the fusion of multiple data streams in the total data buffer at the receiving end. Subsequently, the fused data in the buffer is sent to the preset sentiment classification model via the sendToAIModel interface of the MCP Client. After the preset sentiment classification model completes sentiment inference based on preset prompts, it returns an inference result in the format {"Inference_result":{"value":"0"}}. This inference result is sent to the vehicle's 3D instance via the sendTo3D interface of the MCP Server. The vehicle's 3D instance parses the value of 0 in the inference result and determines that the current user's sentiment is "joy" according to the preset mapping relationship. Then, it calls the material parameter set corresponding to the mood in the 3D engine for real-time rendering and projects the rendered 3D scene onto the vehicle's display screen to realize the display of a vehicle's 3D mood wallpaper that matches the user's sentiment. By pre-setting the mapping relationship between emotion categories and material parameter sets, as well as basic 2D and / or 3D scene models locally in the vehicle system, the corresponding material parameters can be directly retrieved for real-time rendering after the cloud inference results are returned. This avoids the delay and computational overhead caused by dynamically generating scene content. At the same time, it supports synchronizing the rendered visual atmosphere to environmental devices such as in-vehicle ambient lighting. Thus, while ensuring real-time response, it achieves consistent adaptation between the in-vehicle visual experience and the user's emotional state, further enhancing the emotional interaction immersion of the smart cockpit.

[0076] Optionally, after generating the image and text information matching the user's emotional information, the method further includes: The graphic information is output through at least one display unit of the vehicle.

[0077] In this optional embodiment, the display unit may include one or more of the following: a display screen type display unit, for example, the vehicle's central control display screen is the main output interface, which can be used to display static images, dynamic wallpapers, videos, or virtual car model renderings; the instrument panel display screen can be used to display concise graphics or ambient prompts that match the user's mood; a head-up display (HUD) can project emotion-matching graphics or text onto the windshield, such as projecting a floating flower icon when the user is in a good mood; for some models with rear entertainment screens or passenger-side screens, these screens can also output graphic and text information synchronously or independently. In addition, in some designs, the vehicle's sunroof or side window glass can be used as a transparent display screen (such as a dimming glass projection), which can also serve as a display unit to output matching graphic and text images.

[0078] Voice-over display units: These units use speech synthesis to output the semantic content of text in an auditory manner, serving as a supplement or alternative to visual displays. For example, if the generated emotion-matching graphic information includes the encouraging phrase "Relax, take a deep breath," the system can simultaneously or separately broadcast this phrase through the car's speakers, allowing users to perceive emotional interaction without looking at the screen. This voice output method can be seen as an auditory representation of the "text" in the graphic information.

[0079] In the lighting-linked display unit, to create a more immersive emotional atmosphere, the vehicle's cabin lighting system can be linked with the output of text and images. For example, when the main color of the text and images is warm orange (corresponding to the emotion of "joy"), the system sends color control commands to the ambient lighting controller via the in-vehicle bus, causing the ambient lights to synchronously display warm orange with a breathing-like brightness change; when the text and images are in a cool blue tone (corresponding to "calm" or "sadness"), the ambient lights adjust to the corresponding cool color temperature and soft lighting effect. Furthermore, reading lights, welcome lights, and even exterior lights (such as daytime running light color changes) can also participate in the linkage. The lighting linkage in this embodiment does not directly output the text and images themselves, but rather expands the range of emotional expression based on the visual style parameters of the text and images, enhancing the consistency of the overall cabin experience.

[0080] Alternatively, multiple display units can be combined for output. In practical applications, multiple display units can be mobilized to work collaboratively simultaneously. For example, the central control screen displays a rendered image (text and graphics) of a 3D virtual car model, while the ambient lighting dims according to the primary color of the material parameters, and a simple emoticon appears on the HUD. This multimodal output method can enhance the user's emotional resonance from multiple dimensions, including visual, auditory, and even ambient light perception. In the specific implementation process, the above display units can be flexibly selected and combined according to the vehicle's hardware configuration and interaction design requirements, which will not be elaborated further here.

[0081] like Figure 2As shown in the figure, an embodiment of this application provides a vehicle display graphic generation device, comprising: The emotion perception and processing unit is used to acquire emotion-related data to be inferred from the target user associated with the vehicle, and to obtain user emotional information based on the data to be inferred. The generation unit generates image and text information that matches the user's emotional information.

[0082] like Figure 3 As shown in the figure, an electronic device provided in this application includes a memory and a processor; The memory is used to store computer programs; The processor is used to implement the above-described method for generating vehicle display images and text when executing the computer program.

[0083] Alternatively, an electronic device includes a memory and a processor coupled to the memory; the memory is configured to store a computer program; the processor is configured to perform the following operations when the computer program is executed: Acquire data related to user emotions to be inferred, and obtain user emotional information based on the data to be inferred; Based on the user's emotional information, a mood wallpaper matching the user's emotional information is generated.

[0084] Electronic devices include a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM can also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0085] The electronic device and the vehicle display graphic generation method provided in this embodiment can produce basically the same technical effects, and will not be described again here.

[0086] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the vehicle display graphic generation method described above.

[0087] Alternatively, a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the following operations: Acquire data related to user emotions to be inferred, and obtain user emotional information based on the data to be inferred; Based on the user's emotional information, a mood wallpaper matching the user's emotional information is generated.

[0088] The computer-readable storage medium provided in this embodiment has essentially the same technical effect as the vehicle display graphic generation method, and will not be described in detail here.

[0089] This application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method for generating vehicle display images and text.

[0090] The computer program product provided in this embodiment has essentially the same technical effect as the vehicle display graphic generation method, and will not be described in detail here.

[0091] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application according to actual needs. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

[0092] Although the above disclosure is provided, the scope of protection of this application is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of this application, and all such changes and modifications will fall within the scope of protection of this application.

Claims

1. A vehicle display image generation method characterized by comprising: include: Obtain the emotion-related data of the target user associated with the vehicle, and obtain the user's emotional information based on the data to be inferred; Generate image and text information that matches the user's emotional information.

2. The vehicle display graphic generation method according to claim 1, characterized in that, The data to be inferred includes data collected by at least two sensors.

3. The vehicle display graphic generation method according to claim 1, characterized in that, The data to be inferred includes at least one type of user sentiment data from the vehicle's sensor data, the user-side ecosystem data corresponding to the vehicle, and the vehicle's external environment data; and / or, The graphic information includes at least one of the following: images, videos, text, live wallpapers, virtual car model renderings, and cockpit lighting-linked visual scenes.

4. The method for generating vehicle display images and text according to claim 1, characterized in that, The data to be inferred includes at least two user emotion data points, and obtaining user emotion information based on the data to be inferred includes: The at least two user sentiment data are fused to obtain the fused data to be inferred; The user's emotional information is obtained by performing reasoning actions based on the data to be fused using a preset emotion classification model.

5. The vehicle display graphic generation method according to claim 4, characterized in that, The step of obtaining the user's emotional information by performing inference actions based on the data to be fused using a preset emotion classification model includes: Using the preset sentiment classification model, based on the task instructions in the preset prompt words, the data to be fused is semantically understood and sentiment analyzed, and the user sentiment information is output according to the output format in the preset prompt words.

6. The method for generating vehicle display images and text according to claim 1, characterized in that, The generation of image and text information matching the user's emotional information includes: The rendering engine's material parameters are adjusted using a set of material parameters corresponding to the user's emotional information, and the adjusted rendering engine is used to generate graphic and textual information.

7. The vehicle display graphic generation method according to claim 6, characterized in that, The step of adjusting the material parameters of the rendering engine using a set of material parameters corresponding to the user's emotional information, and then using the adjusted rendering engine to generate graphic and textual information, includes: Based on the mapping relationship between each emotion category in the preset emotion category set and the preset material parameters in the rendering engine, the set of material parameters corresponding to the user's emotion information is determined. The material parameters of the rendering engine are adjusted using the material parameter set to obtain the adjusted rendering engine. The adjusted rendering engine performs real-time rendering based on a preset scene model to generate graphic and textual information that matches the user's emotional information.

8. The method for generating vehicle display images and text according to any one of claims 1 to 7, characterized in that, After generating the image and text information matching the user's emotional information, the method further includes: The graphic information is output through at least one display unit of the vehicle.

9. An electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement the vehicle display graphic generation method as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when read and run by a processor, implements the vehicle display graphic generation method as described in any one of claims 1 to 8.

11. A computer program product, characterized in that, The system includes a computer program that, when executed by a processor, implements the vehicle display graphic generation method as described in any one of claims 1 to 8.

Images