CN-116385856-B - Data transmission method, device and storage medium

CN116385856BCN 116385856 BCN116385856 BCN 116385856BCN-116385856-B

Abstract

The invention belongs to the field of data transmission, and discloses a data transmission method, equipment and a storage medium. The method comprises the steps of extracting visual features of a target picture, projecting the visual features to a potential low-dimensional space to obtain potential low-dimensional features, projecting the potential low-dimensional features to a semantic feature space to obtain semantic features based on a multi-level semantic encoder, performing category judgment on the semantic features based on a first semantic knowledge base to obtain semantic errors, determining the target features of the target picture according to the semantic errors, and sending the target features to data receiving equipment. The method comprises the steps of extracting visual characteristics of a target picture, determining the target characteristics of the target picture according to semantic errors corresponding to the visual characteristics, and sending the target characteristics to data receiving equipment. Compared with the existing mode of directly sending the target picture to the data receiving equipment, the mode can reduce transmission delay and save transmission bandwidth.

Inventors

SUN YAPING
CHEN HAO
XU XIAODONG
CUI SHUGUANG
ZHANG PING

Assignees

鹏城实验室
香港中文大学（深圳）

Dates

Publication Date: 20260505
Application Date: 20230322

Claims (8)

1. A data transmission method, characterized by being applied to a data transmission apparatus, comprising the steps of: extracting visual features of a target picture, and projecting the visual features to a potential low-dimensional space to obtain potential low-dimensional features; Based on a multi-level semantic encoder, projecting the potential low-dimensional features to a semantic feature space to obtain semantic features; Performing category judgment on the semantic features based on a first semantic knowledge base to obtain semantic errors; Determining target characteristics of the target picture according to the semantic errors, and sending the target characteristics to data receiving equipment; The step of determining the target feature of the target picture according to the semantic error and sending the target feature to a data receiving device comprises the following steps: when the semantic error is smaller than a preset threshold, taking estimated category information corresponding to the semantic error as target characteristics of the target picture, and sending the target characteristics to data receiving equipment; And when the semantic error is greater than or equal to the preset threshold, taking the semantic feature as a target feature of the target picture, and sending the target feature to the data receiving equipment.
2. The data transmission method according to claim 1, wherein the step of taking the semantic feature as a target feature of the target picture and transmitting the target feature to the data receiving device when the semantic error is greater than or equal to the preset threshold value further comprises: Receiving semantic feature class judgment results fed back by the data receiving equipment; and when the semantic error of the receiving end in the semantic feature type judgment result is larger than the preset threshold, transmitting the visual features of the target picture to the data receiving equipment.
3. The data transmission method according to claim 1, wherein the estimated category information includes estimated category identifications and/or estimated category semantic vectors; and when the semantic error is smaller than a preset threshold, taking the estimated category information corresponding to the semantic error as a target feature of the target picture, and sending the target feature to a data receiving device, wherein the method comprises the following steps of: When the semantic error is smaller than a preset threshold, the estimated category identification is used as a target feature of the target picture, and the target feature is sent to data receiving equipment; receiving an identification result fed back by the data receiving equipment; And when the identification result is that the estimated category semantic vector corresponding to the estimated category identification does not exist in the second semantic knowledge base of the data receiving equipment, the estimated category semantic vector is used as the target feature of the target picture to be sent to the data receiving equipment.
4. A data transmission method applied to the data transmission method according to any one of claims 1 to 3, characterized in that it is applied to a data receiving apparatus, the data transmission method comprising the steps of: receiving target characteristics of a target picture sent by data sending equipment; based on a second semantic knowledge base and/or a multi-level semantic decoder, performing category judgment on the target features, and identifying the target pictures; the step of identifying the target picture based on the category judgment of the target features by the second semantic knowledge base and/or the multi-level semantic decoder comprises the following steps: When the target feature is an estimated category identifier, judging whether an estimated category semantic vector corresponding to the estimated category identifier exists in a second semantic knowledge base or not based on the second semantic knowledge base; If so, acquiring the estimated category semantic vector from the second semantic knowledge base according to the estimated category identifier, and identifying the target picture according to the estimated category semantic vector.
5. The data transmission method according to claim 4, wherein the step of identifying the target picture based on the category decision of the target feature by the second semantic knowledge base and/or the multi-level semantic decoder comprises: when the target feature is a potential low-dimensional feature, projecting the potential low-dimensional feature to a semantic feature space according to a multi-level semantic decoder to obtain the target semantic feature; performing category judgment on the target semantic features based on the second semantic knowledge base to obtain estimated category information; and identifying the target picture according to the estimated category information.
6. A data transmission device, characterized in that the device comprises a memory, a processor and a data transmission program stored on the memory and executable on the processor, the data transmission program being configured to implement the steps of the data transmission method according to any one of claims 1 to 3.
7. A data receiving device, characterized in that the device comprises a memory, a processor and a data transmission program stored on the memory and executable on the processor, the data transmission program being configured to implement the steps of the data transmission method according to any one of claims 4 to 5.
8. A storage medium having stored thereon a data transmission program which, when executed by a processor, implements the steps of the data transmission method according to any one of claims 1 to 3 or 4 to 5.

Description

Data transmission method, device and storage medium Technical Field The present invention relates to the field of data transmission technologies, and in particular, to a data transmission method, device, and storage medium. Background Remote zero sample semantic understanding difficulties exist widely in future 6G typical intelligent scenarios. For example, in an automatic driving scenario, an unmanned vehicle often needs to rely on wireless communication with a remote vehicle or a base station to acquire a remote traffic condition in advance, so as to optimally plan a route. However, there are often new situations in road traffic, such as emergency accidents, new vehicles on new roads, and the like. These scene pictures are often not available in advance, are not visible during the training process, belong to an unknown class, and therefore automatic vehicles face the difficult problem of remote zero sample image recognition. Aiming at the problem of zero sample image recognition, the traditional recognition method often depends on a training data set of large-scale strong supervision labeling or directly sends an original picture, is not suitable for solving the problem of zero sample, consumes a great amount of manpower, material resources and time, and is inconsistent with the requirements of intelligent and simple communication on low time delay, light weight and high efficiency. The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present invention and is not intended to represent an admission that the foregoing is prior art. Disclosure of Invention The invention mainly aims to provide a data transmission method, equipment and a storage medium, and aims to solve the technical problem that a remote zero sample image cannot be identified in the prior art. To achieve the above object, the present invention provides a data transmission method applied to a data transmission apparatus, the method comprising the steps of: extracting visual features of a target picture, and projecting the visual features to a potential low-dimensional space to obtain potential low-dimensional features; Based on a multi-level semantic encoder, projecting the potential low-dimensional features to a semantic feature space to obtain semantic features; Performing category judgment on the semantic features based on a first semantic knowledge base to obtain semantic errors; And determining target characteristics of the target picture according to the semantic error, and sending the target characteristics to data receiving equipment. Optionally, the step of determining the target feature of the target picture according to the semantic error and sending the target feature to a data receiving device includes: when the semantic error is smaller than a preset threshold, taking estimated category information corresponding to the semantic error as target characteristics of the target picture, and sending the target characteristics to data receiving equipment; And when the semantic error is greater than or equal to the preset threshold, taking the semantic feature as a target feature of the target picture, and sending the target feature to the data receiving equipment. Optionally, after the step of using the semantic feature as the target feature of the target picture and sending the target feature to the data receiving device when the semantic error is greater than or equal to the preset threshold, the method further includes: Receiving semantic feature class judgment results fed back by the data receiving equipment; and when the semantic error of the receiving end in the semantic feature type judgment result is larger than the preset threshold, transmitting the visual features of the target picture to the data receiving equipment. Optionally, the estimated category information includes an estimated category identification and/or an estimated category semantic vector; and when the semantic error is smaller than a preset threshold, taking the estimated category information corresponding to the semantic error as a target feature of the target picture, and sending the target feature to a data receiving device, wherein the method comprises the following steps of: When the semantic error is smaller than a preset threshold, the estimated category identification is used as a target feature of the target picture, and the target feature is sent to data receiving equipment; receiving an identification result fed back by the data receiving equipment; And when the identification result is that the estimated category semantic vector corresponding to the estimated category identification does not exist in the second semantic knowledge base of the data receiving equipment, the estimated category semantic vector is used as the target feature of the target picture to be sent to the data receiving equipment. In addition, to achieve the above object, the present invention also provides a data transm