CN-122002048-A - Communication method and related device

Abstract

The application provides a communication method and a related device, which are used for updating a model for processing image or video data and reducing the generated communication loss between a transmitting end and a receiving end. In the method, after the communication device determines the model update information to the opposite terminal, the target model is updated at a proper time based on the model update time indicated by the model update information, so that the target model can be updated at a proper time after being successfully configured, the problem that the performance of the model is reduced due to switching of application scenes or updating of image definition requirements is avoided, efficient transmission of image data or video data between the communication devices is ensured, and better user experience is improved when a user browses images or views videos.

Inventors

• XU RUI
• QIN YI
• YANG YUWEN
• QI HONG

Assignees

• 华为技术有限公司

Dates

Publication Date

20260508

Application Date

20241105

Claims (17)

1. 1. A method of communication, comprising: Determining model update information, the model update information indicating a model update opportunity; And updating a target model based on the model updating information, wherein the target model comprises an information source compression model or a data recovery model, the information source compression model is used for carrying out feature extraction or compression on video data or image data to obtain data to be recovered, and the data recovery model is used for carrying out feature recovery or decompression on the data to be recovered to obtain the video data or the image data.
2. 2. The method of claim 1, wherein determining model update information comprises: Determining a target duration, wherein the model updating information comprises the target duration; the updating the target model based on the model update information includes: updating the target model at a first moment, wherein the time interval between the first moment and a second moment is the target duration, and the second moment is the updating moment of the target model.
3. 3. The method of claim 2, wherein before updating the target model at the first time, the method further comprises: Determining target data set information, wherein the target data set information comprises a target data set or at least one of identity marks of the target data set; The updating the target model at the first time includes: The target model is updated at a first time using the target dataset information.
4. 4. The method of claim 3, wherein the target model is the data recovery model, the method further comprising: and sending the target duration and the target data set information.
5. 5. The method of claim 2, wherein the target model is the source compression model, and wherein prior to the determining the target time period, the method further comprises: transmitting a reference time length, wherein the reference time length comprises the target time length; The determining the target duration includes: and receiving the target duration.
6. 6. The method of claim 2, wherein the target model is the source compression model, and wherein prior to the determining the target time period, the method further comprises: Receiving a reference time length, wherein the reference time length comprises the target time length; The determining the target duration includes: Selecting a target duration from the reference durations; The method further comprises the steps of: and sending the target duration.
7. 7. The method of claim 1, wherein determining model update information comprises: Determining a model update signal, the model update information comprising the model update signal; The updating the target model based on the model update information includes: After determining the model update signal, the target model is updated.
8. 8. The method of claim 7, wherein the model update signal is a signal generated when a distance between a first feature vector and a second feature vector is greater than a first threshold, the first feature vector and the second feature vector being a result of feature extraction of video data or image data by the source compression model in adjacent time periods.
9. 9. The method of claim 7, wherein the model update signal is a signal generated when a first degree of impairment or a second degree of impairment is greater than a second threshold, the first degree of impairment being a degree of impairment of image data generated by the data recovery model from the data to be recovered, the second degree of impairment being a degree of impairment of video data generated by the data recovery model from the data to be recovered.
10. 10. The method of claim 7, wherein the model update signal is a signal generated when a third degree of impairment is greater than a third threshold, the third degree of impairment being determined based on a similarity of a first image obtained by recovering the first string using the data recovery model and a second image being an image having a correspondence with the first string, the second image being similar to an image to which the data to be recovered corresponds.
11. 11. The method of claim 10, wherein the target model is the source compression model, and wherein determining a model update signal comprises: receiving the model update signal; before determining the model update signal, the method further comprises: and transmitting the first character string or the first character string and the second image.
12. 12. The method according to claims 8 to 11, wherein determining a model update signal comprises: Generating the model update signal; after determining the model update signal, the method further comprises: And sending the model update signal.
13. 13. The method according to any one of claims 1 to 11, wherein before determining the model update information, the method further comprises: Determining a model to be trained and a preset data set; And training the model to be trained by using the preset data set to obtain the target model.
14. 14. A communication device, comprising: a processing unit configured to determine model update information indicating a model update opportunity; The model training unit is used for updating a target model based on the model updating information, wherein the target model comprises an information source compression model or a data recovery model, the information source compression model is used for carrying out feature extraction or compression on video data or image data to obtain data to be recovered, and the data recovery model is used for carrying out feature recovery or decompression on the data to be recovered to obtain the video data or the image data.
15. 15. A communications apparatus comprising a processor coupled to a memory; Stored in the memory are instructions which, when executed on the processor, cause the communication device to carry out the method of any one of claims 1 to 13.
16. 16. A computer readable storage medium storing instructions which, when run on a processor, cause the method of any one of claims 1 to 13 to be implemented.
17. 17. A computer program product, characterized in that the computer program product, when executed on a computer, causes the method of any one of claims 1 to 13 to be implemented.

Description

Communication method and related device Technical Field The present application relates to the field of communications, and in particular, to a communication method and a related device. Background With the continuous updating of streaming media, higher definition image quality and faster transmission rate become main development directions of data transmission of streaming media. Generated artificial intelligence (generation ARTIFICIAL INTELLIGENCE, generationAI) provides a technological basis for more efficient transmission of data. And the characteristic extraction capacity and the characteristic recovery capacity of the artificial intelligence are utilized, at the transmitting end of the audio and video data, after the audio and video data are subjected to characteristic extraction or compression by using the model, the characteristic extraction result or the compression result is transmitted to the receiving end, and the receiving end recovers the audio and video data according to the characteristic extraction result or the compression result. However, once the model under the scene is configured successfully, the model is not updated any more, and a large span may appear in the image or video data transmitted between the sending end and the receiving end, so that when the picture type is switched, performance degradation caused by that the model cannot adapt to the switching of the picture type occurs, and the generated communication loss is high in the transmission process of the image or the video. Disclosure of Invention The application provides a communication method and a related device, which are used for updating a model for processing image or video data and reducing the generated communication loss between a transmitting end and a receiving end. The first aspect of the present application provides a communication method, which is performed by a communication device, where the communication device may be a transmitting end (e.g. a user equipment), a receiving end (e.g. a base station or a server), an entity in the transmitting end (e.g. a processor, a chip or a chip system, etc.), an entity in the receiving end (e.g. a processor, a chip or a chip system, etc.), or the method may also be implemented by a logic module or software capable of implementing all or part of the functions of the transmitting end or the functions of the receiving end. In the first aspect and its possible implementation manner, the method is described as an example executed by the transmitting end or the receiving end. In the method, a transmitting end or a receiving end determines model update information and updates a target model at a proper timing based on the model update information, wherein the model update information indicates a model update timing. Based on the scheme, after the communication device determines the model updating information to the opposite terminal, the target model is updated at a proper time based on the model updating time indicated by the model updating information, so that the target model can be updated at a proper time after being successfully configured, the situation that the application scene is switched or the model performance is reduced due to the updating of the image definition requirement is avoided, the efficient transmission of image data or video data between the communication devices is ensured, and better user experience is obtained when a user browses images or views videos. In an implementation example, the model update information may include a target time length (period), and the communication device updates the target model at a first time after determining the target time length, where a time interval between the first time and a second time is the target time length, and the second time is an update time of the target model, and the second time is before the first time. That is, the communication device performs updating of the target model by determining the time interval between two model updating timings. It should be understood that the update time of the target model is the time when the model is trained to obtain the target model. In a possible implementation manner of the first aspect, in a case of updating the target model according to the target duration, a target data set used for updating the target model may be preset, and the determination of the target data set is implemented by determining target data set information, where the target data set information is at least one of a target data set or an identity of the target data set. And updating the target model at the first time using the target dataset information based on the target duration. Based on the scheme, the communication device can send the target time length to the opposite terminal equipment and simultaneously send the target data set information to the opposite terminal equipment, so that the communication device and the opposite terminal equipment can update the model based on the same or