Multi-level feature multicast transmission method and device, equipment, storage medium and product

Abstract

The invention discloses a multi-level feature multicast transmission method and device, equipment, a storage medium and a product, and relates to the technical field of semantic communication.The method comprises the steps that multi-level features are obtained according to a target picture, the multi-level features comprise semantic features, semantic errors of different levels are determined according to the semantic features, and the target picture is obtained; and determining the communication resource of the data receiving device in the current hierarchy according to the target feature of the target picture, thereby obtaining the receiving capabilities of different users, and facilitating reasonable resource allocation during subsequent semantic communication for multiple users. An optimization problem is established based on communication resources of different levels of data receiving equipment and semantic errors of different levels to obtain an optimal solution, and furthermore, multicast transmission is performed on target features based on the optimal solution, so that the communication resource overhead in a multi-user scene is reduced, the requirements of different mobile agents are met in information transmission, and the information transmission efficiency is improved. And the transmission time delay and the resource consumption are reduced on the whole.

Description

Technical Field

[0001] The present application relates to the technical field, and in particular to a multi-level feature multicast transmission method, apparatus, device, storage medium and product. Background Art

[0002] With the rapid development of mobile communications and Internet technologies, the demand for high-speed, low-latency wireless access has increased dramatically, and traditional communication systems have approached the Shannon theoretical limit. In the future typical 6G intelligent scenarios, there are widespread problems with remote zero-sample semantic understanding. For example, unmanned vehicles in autonomous driving scenarios often need to rely on wireless communications with distant vehicles or base stations to obtain distant traffic conditions in advance, so as to optimally plan routes; however, road traffic often has new situations such as emergency accidents and new vehicles on the road. These scene images are often not available in advance, are invisible during training, and belong to unknown categories, that is, autonomous vehicles face the problem of remote zero-sample image recognition.

[0003] Currently, when a server transmits features to multiple mobile agents, there is a large communication resource overhead.

[0004] Therefore, how to reduce the communication resource overhead in multi-user scenarios is a problem that needs to be solved urgently. Summary of the invention

[0005] The main purpose of the present application is to provide a multi-level feature multicast transmission method, device, equipment, storage medium and product, aiming to solve the technical problem of high communication resource overhead in multi-user scenarios.

[0006] To achieve the above objectives, the present application proposes a multi-level feature multicast transmission method, the method comprising: Acquire multi-level features according to the target image, wherein the multi-level features include semantic features; Determining semantic errors at different levels according to the semantic features, and determining communication resources of a data receiving device at a current level according to target features of the target image; Based on the communication resources of data receiving devices at different levels and the semantic errors at the different levels, an optimization problem is established to obtain an optimal solution; The target features are multicast transmitted based on the optimal solution.

[0007] In one embodiment, the step of establishing an optimization problem based on the semantic errors at different levels and the communication resources of the data receiving devices at different levels to obtain an optimal solution includes: Determining semantic losses corresponding to different levels according to the semantic errors of the different levels; Determining multicast delays corresponding to different levels according to communication resources of data receiving devices in the different levels; An optimization problem is established according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels to obtain an optimal solution.

[0008] In one embodiment, the step of establishing an optimization problem according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels to obtain an optimal solution includes: Establishing a multi-level multicast selection strategy, a bandwidth allocation strategy, and a power allocation strategy according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels; By optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy, the semantic loss and the average maximum multicast delay are minimized to obtain an optimal solution.

[0009] In one embodiment, the step of obtaining the optimal solution by optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy to minimize semantic loss and average maximum multicast delay includes: Under any given multicast grouping strategy, the corresponding bandwidth allocation strategy and the corresponding power allocation strategy are jointly optimized to obtain the optimal bandwidth allocation strategy and the optimal power allocation strategy; Based on the optimal bandwidth allocation strategy and the optimal power allocation strategy, an optimal strategy for multicast grouping is obtained.

[0010] In one embodiment, the step of obtaining the optimal strategy for multicast grouping based on the optimal bandwidth allocation strategy and the optimal power allocation strategy includes: Based on the optimal bandwidth allocation strategy and the optimal power allocation strategy, generating all multicast grouping strategies; The minimum value of the semantic loss and the average maximum multicast delay is searched among all the multicast grouping strategies, and the multicast grouping strategy corresponding to the minimum value is the multicast grouping optimal strategy.

[0011] In one embodiment, the multi-level features include visual features, potential low-dimensional intermediate features, and semantic features, and the step of acquiring the multi-level features according to the target image includes: Extracting visual features of the target image, and projecting the visual features into a potential low-dimensional intermediate space to obtain the potential low-dimensional intermediate features; Based on a multi-level semantic encoder, the potential low-dimensional intermediate features are projected into a semantic feature space to obtain the semantic features.

[0012] In addition, to achieve the above purpose, the present application also proposes a multi-level feature multicast transmission device, the multi-level feature multicast transmission device comprising: A feature acquisition module, used to acquire multi-level features according to a target image, wherein the multi-level features include semantic features; A transmission resource module, used to determine the semantic errors of different levels according to the semantic features, and determine the communication resources of the data receiving device in the current level according to the target features of the target picture; A transmission optimization module, used to establish an optimization problem based on the semantic errors at different levels and the communication resources of data receiving devices at different levels to obtain an optimal solution; A multicast transmission module is used to perform multicast transmission on the target feature based on the optimal solution.

[0013] In addition, to achieve the above-mentioned objectives, the present application also proposes a multi-level feature multicast transmission device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the multi-level feature multicast transmission method as described above.

[0014] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the multi-level feature multicast transmission method described above are implemented.

[0015] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the multi-level feature multicast transmission method as described above are implemented.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: According to the target image, multi-level features are obtained, including semantic features. By determining the semantic errors of different levels according to the semantic features, and determining the communication resources of the data receiving device in the current level according to the target features of the target image, the receiving capabilities of different users are obtained, which is convenient for reasonable resource allocation in the subsequent semantic communication for multiple users. By establishing an optimization problem based on the communication resources of data receiving devices at different levels and the semantic errors at different levels, the optimal solution is obtained. Furthermore, the target features are multicasted based on the optimal solution, which reduces the communication resource overhead in the multi-user scenario. In information transmission, the needs of different mobile agents are met, and the transmission delay and resource consumption are reduced overall, providing strong technical support for the realization of intelligent and simple communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 This is a schematic diagram of the MEC system of the embodiment of the present application; Figure 2 This is a flowchart of the first embodiment of the multi-level feature multicast transmission method of the present application; Figure 3 This is a flowchart of a second embodiment of the multi-level feature multicast transmission method of the present application; Figure 4 A schematic diagram of constructing a multi-level feature extractor for a multi-level feature according to an embodiment of the present application; Figure 5 This is a schematic diagram of the module structure of a multi-level feature multicast transmission device according to an embodiment of the present application; Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the multi-level feature multicast transmission method in the embodiment of the present application.

[0020] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0022] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0023] This application proposes a new edge server (Multi-Access Edge Computing, MEC) system, such as Figure 1 As shown in Figure 1, the MEC system consists of a single MEC server and multiple mobile intelligent agents, each of which is endowed with a single multi-level feature extractor and a semantic knowledge base (Semantic Knowledge Base, SKB). is the number of mobile agents, and Represented as a collection of mobile agents. Both the MEC server and the mobile agent have a semantic knowledge base (SKB) and a multi-level semantic codec. The SKB is defined as a collection of semantic vectors corresponding to some image categories, and the multi-level semantic codec describes the mapping relationship between the dimension size, classification accuracy, and semantic loss of different levels.

[0024] Taking image transmission as an example, assuming that the MEC server perceives a certain image sample and sends it to the mobile agent, the mobile agent needs to identify the specific semantic description of the image, which is called a remote image recognition task. Since the resource conditions such as the size of the semantic knowledge base equipped by the MEC server and different mobile agents may be different, it is impossible to ensure that the information sent by MEC to the mobile agent every time is estimated category information with a very high compression rate. Therefore, research is carried out on multi-level feature extraction, transmission and resource allocation problems for multi-user transmission. This application mainly focuses on downlink transmission in multi-user scenarios.

[0025] Based on this, the embodiment of the present application provides a multi-level feature multicast transmission method, referring to Figure 2 , Figure 2 This is a flowchart of the first embodiment of the multi-level feature multicast transmission method of the present application.

[0026] In this embodiment, the multi-level feature multicast transmission method includes steps S10 to S40: Step S10, obtaining multi-level features according to the target image.

[0027] It should be noted that the target image can be understood as the image input into the MEC system, and the multi-level features can be understood as the characteristics of the image at different processing stages or different abstract levels. Exemplarily, a multi-feature extractor can be used to obtain the multi-level features of the target image.

[0028] Step S20, determining semantic errors of different levels according to the multi-level features, and determining communication resources of the data receiving device in the current level according to the target features of the target image.

[0029] It should be noted that semantic error can be understood as the loss of image information caused by compression, noise and other factors during the transmission or processing process, especially the loss of key information that affects the understanding of the image content. The data receiving device can be multiple mobile intelligent agents, and the communication resources of the data receiving device can be the network resources required for the data during the transmission process, including but not limited to bandwidth, delay, transmission power, etc.

[0030] Step S30, based on the communication resources of data receiving devices at different levels and semantic errors at different levels, an optimization problem is established to obtain an optimal solution.

[0031] Exemplarily, the communication resources of data receiving devices at different levels may be optimized while minimizing the semantic error, and a corresponding objective function may be established so that the objective function reaches a minimum (or maximum) solution.

[0032] Step S40: multicast the target features based on the optimal solution.

[0033] It should be noted that the target feature, i.e., a feature of the image sample finally determined to be sent by the MEC, is at least one of the estimated category of the target image, the semantic vector corresponding to the minimum semantic error with the semantic feature in the first semantic knowledge base, the semantic feature of the target image, the potential low-dimensional intermediate feature, and the visual feature. Multicast transmission is a communication method for effectively transmitting data in a computer network, which allows a sender to send data to a specific group of receivers at the same time.

[0034] In this embodiment, multi-level features are obtained according to the target image, and the multi-level features include semantic features. By determining the semantic errors of different levels according to the semantic features, and determining the communication resources of the data receiving device in the current level according to the target features of the target image, the receiving capabilities of different users are obtained, which is convenient for reasonable resource allocation in the subsequent semantic communication for multiple users. By establishing an optimization problem based on the communication resources of data receiving devices at different levels and the semantic errors at different levels, the optimal solution is obtained. Furthermore, the target features are multicasted based on the optimal solution, which reduces the communication resource overhead in the multi-user scenario. In information transmission, the needs of different mobile agents are met, and the transmission delay and resource consumption are reduced overall, providing strong technical support for the realization of intelligent and simple communication.

[0035] Reference Figure 3 , Figure 3 This is a flow chart of the second embodiment of the multi-level feature multicast transmission method of the present application, based on the above Figure 2 The first embodiment shown proposes a second embodiment of the multi-level feature multicast transmission method of the present application.

[0036] In the second embodiment, the step S10 includes: Step S101, extracting visual features of the target image, and projecting the visual features to a potential low-dimensional intermediate space to obtain potential low-dimensional intermediate features.

[0037] It should be noted that visual features can be attributes automatically extracted from images that can describe the content of the image, such as the color, texture, shape, edge, etc. of the image. The potential low-dimensional intermediate space can be a mathematical space with a dimension lower than the original data space, which is used to store the feature representation after dimensionality reduction. The potential low-dimensional intermediate feature can be understood as the feature representation obtained after dimensionality reduction, which encodes the key information of the original image in a lower dimension.

[0038] Step S102: Based on the multi-level semantic encoder, the potential low-dimensional intermediate features are projected into the semantic feature space to obtain semantic features.

[0039] It should be noted that the multi-level semantic encoder can be understood as a table model structure, which usually contains multiple levels or layers, each of which is responsible for extracting semantic information at different levels. The semantic feature space is used to further process the potential low-dimensional intermediate feature information. Semantic features focus more on describing the abstract and high-level meaning of the image content, such as the category, attributes, and relationship of the object.

[0040] Based on visual and semantic autoencoders (AE), given any image sample ,The multi-level feature extractor adopted is as follows.

[0041] ① Level 1, visual features: A pre-trained large-scale CNN (e.g., GoogleNet, ResNet, etc.) is used to project the input image samples to visual features, i.e., ,in represents a pre-trained large-scale CNN, Represents the dimension of visual features.

[0042] ②Level 2, potential low-dimensional intermediate features: The visual encoder projects the visual features into a space of latent low-dimensional intermediate features, i.e., ,in represents the visual encoder, Represents the dimension of the intermediate features.

[0043] ③Level 3, semantic features: The semantic decoder projects the latent low-dimensional intermediate features into the space of semantic features, i.e., ,in represents the semantic decoder, The dimension representing the semantic features.

[0044] ④Level 4, estimated category labels: By measuring the projected semantic representation The semantic similarity with the category prototype in SKB is used to estimate the category of the sample, that is, ,in, Indicates that there is a category set (category samples seen during training), Indicates category The semantic vector of .

[0045] In this embodiment, by extracting the visual features of the target image, the visual information of the target image can be obtained, and by projecting the visual features to the potential low-dimensional intermediate space, the dimension of the feature vector is reduced, the computational efficiency is improved, and the potential low-dimensional intermediate features are obtained. Furthermore, based on the multi-level semantic encoder, the potential low-dimensional intermediate features are projected to the semantic feature space to obtain semantic features, which can capture and encode the complex semantic information in the image, so that the image content can be understood and analyzed more accurately, thereby providing users with more personalized and intelligent services.

[0046] Reference Figure 4 , Figure 4 This is a flow chart of the third embodiment of the multi-level feature multicast transmission method of the present application. Figure 3 The second embodiment shown proposes the third embodiment of the multi-level feature multicast transmission method of the present application.

[0047] In the third embodiment, the step S30 includes: Step S301 , determining semantic losses corresponding to different levels according to semantic errors at different levels.

[0048] For example, let Indicates k Is the user in l Layer multicast group.

[0049] For level 1, visual feature multicast, the semantic loss can be expressed as: when When the sample n The visual feature vector (expressed as ) Multicast from MEC server to the collection Mobile agent in (at this time , i.e., the set of all agents that use level 1 transmission). Then, each agent Based on its multi-level feature extractor and its SKB, the category is estimated as formula (1):

[0050] In formula (1), represents the category estimated by each agent at level 1 based on its multi-level feature extractor and its SKB, Indicates k The user's knowledge base, Indicates the corresponding c The semantic knowledge vector of each category, Indicates k Semantic decoder for each user, Indicates k The visual encoder of each user, Indicatesn The visual feature vector of samples, F represents the Frobenius norm, n Indicates n samples, k Indicates k users.

[0051] acting The corresponding semantic loss is expressed as formula (2):

[0052] in, Representation Agent The corresponding semantic loss, in the set The maximum semantic loss incurred by a mobile agent within the range is expressed as formula (3)

[0053] in, Indicates in the collection The maximum semantic loss incurred in the mobile agent within the range, for the second level, latent low-dimensional intermediate feature multicast, the semantic loss can be expressed as: when When the sample The intermediate features (expressed as ) Multicast from MEC server to the collection Mobile agent in (at this time , i.e., the set of all agents that use level 2 transmission). Then, each agent The estimated category based on its semantic decoder is expressed as formula (4):

[0054] In formula (4), represents the category estimated by each agent at level 2 based on its multi-level feature extractor and its SKB, Indicates k The user's knowledge base, Indicates the corresponding c The semantic knowledge vector of each category, Indicates k Semantic decoder for each user, represents the visual encoder of the kth user, Indicates n The visual feature vector of samples, F represents the Frobenius norm, n Indicates n samples, k Indicates k users.

[0055] The corresponding semantic loss is expressed as formula (5):

[0056] in, Representation Agent The corresponding semantic loss, in the set The maximum semantic loss incurred by a mobile agent within the range is expressed as formula (6):

[0057] in, Indicates in the collection The maximum semantic loss that occurs among mobile agents within the range.

[0058] For level 3, the extracted semantic features (directly obtained from the image) are multicasted, and the semantic loss can be expressed as: when When the sample The semantic feature vector of ) Multicast from MEC server to the collection Then, each agent The estimated category based on its SKB is expressed as formula (7):

[0059] In formula (7), represents the category estimated by each agent at level 3 based on its multi-level feature extractor and its SKB, Indicates k The user's knowledge base, Indicates the corresponding c The semantic knowledge vector of each category, Indicates the corresponding m The semantic knowledge vector of each category, the corresponding semantic loss is expressed as formula (8):

[0060] in, Representation Agent The corresponding semantic loss, Indicates the corresponding c The semantic knowledge vector of each category, Indicates M Semantic decoder for each user, Indicates M The visual encoder of each user, Indicates n The visual feature vector of samples, F represents the Frobenius norm, n Indicatesn samples, M Indicates M users.

[0061] For level 4, the estimated semantic features (the semantic features with the highest similarity to the semantic features directly output from the image found from the sender SKB) are multicast, and the semantic loss can be expressed as: when When , the MEC server first estimates the category of the image sample, that is, formula (9):

[0062] In formula (9), Indicates the category estimated by each agent at level 4 based on its multi-level feature extractor and its SKB, and then the corresponding semantic vector Multicast to collection The corresponding semantic loss is formula (10):

[0063] in, Representation Agent The corresponding semantic loss, in the set The maximum semantic loss incurred by a mobile agent within range is given by Given.

[0064] For level 5, the estimated category labels are multicast and the semantic loss can be expressed as: when When , the MEC server first estimates the category of the image sample, that is, formula (9), and then uses the estimated category label Multicast to collection In order to ensure that the corresponding semantic features are stored in the SKB of the mobile agent, it is necessary to satisfy , The corresponding semantic loss is the same as that of level 4, expressed as formula (11):

[0065] in, Representation Agent The corresponding semantic loss, in the set The maximum semantic loss incurred by a mobile agent within range is given by Given.

[0066] Step S302: determining the multicast delays corresponding to different levels according to the communication resources of the data receiving devices in different levels.

[0067] Exemplarily, based on step S301, the multicast delays corresponding to different levels are expressed as follows. The multicast delay caused by the first level transmission (expressed as ), the first level multicast delay calculation formula is expressed as formula (12):

[0068] in, Indicates the first level multicast delay, represents the visual feature dimension, Q Indicates the number of quantized bits for each dimension of data and the transmission rate It is expressed as formula (13):

[0069] In formula (13), Indicates the first level transmission rate, Indicates the allocation of layer l multicast bandwidth. Indicates the allocation of layer l multicast power, Indicates from MEC server to proxy The channel power gain, represents the noise power spectral density.

[0070] The multicast delay caused by the second level transmission (expressed as ), the second level multicast delay calculation formula is expressed as formula (14):

[0071] in, Indicates the second level multicast delay, represents the potential feature dimension, Q Indicates the number of quantized bits for each dimension of data and the transmission rate It is expressed as formula (15):

[0072] in, Indicates the second level transmission rate, Indicates the allocation of Layer 2 multicast bandwidth. represents the allocated layer 2 multicast power, and the multicast delay caused by the third level transmission (expressed as ), the third level multicast delay calculation formula is expressed as formula (16):

[0073] In formula (16), Indicates the third level multicast delay, Represents semantic feature dimension, transmission rate It is expressed as formula (17):

[0074] in, Indicates the third level transmission rate, Indicates the allocation of Layer 3 multicast bandwidth. The multicast delay caused by the fourth level transmission (expressed as ), the calculation formula for the 4th level multicast delay is expressed as formula (18):

[0075] In formula (18), Indicates the 4th level multicast delay and transmission rate It is expressed as formula (19):

[0076] The multicast delay caused by the fifth level transmission (expressed as ), the fifth level multicast delay calculation formula is expressed as formula (20):

[0077] In formula (20), Indicates the 5th level multicast delay and transmission rate It is expressed as formula (21):

[0078] in, Indicates the 5th level transmission rate, Indicates the allocation of Layer 5 multicast bandwidth. Indicates the allocation of Layer 5 multicast power. Indicates from MEC server to proxy The channel power gain, represents the noise power spectral density.

[0079] Step S303: Establish an optimization problem according to the semantic losses corresponding to different levels and the multicast delays corresponding to different levels to obtain an optimal solution.

[0080] It should be noted that step S303 includes: establishing a multi-level multicast selection strategy, a bandwidth allocation strategy, and a power allocation strategy according to the semantic losses corresponding to different levels and the multicast delays corresponding to different levels; minimizing the semantic losses and the average maximum multicast delay by optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy to obtain the optimal solution.

[0081] Exemplarily, according to step S301 and step S302, it can be concluded that the average semantic loss of the entire MEC system is expressed as formula (22):

[0082] In formula (22), N represents the number of samples, Indicates k User transmission n The first l The semantic error caused by the level features, Indicates k Is the user in l The indicator variable in the multicast user group of the level, if , indicating the k User in l level of multicast user groups, otherwise, The average maximum multicast delay consumption is expressed as formula (23):

[0083] In formula (23), represents the feature dimension, Q Indicates the number of quantization bits for each dimension of data, It is expressed as the downlink rate, which is specifically expressed as formula (24):

[0084] in, Indicates the allocation l Layer multicast bandwidth, Indicates the allocation l Layer multicast power, Indicates k The channel gain of each user is is the noise power spectral density.

[0085] In formula (23), , , ,and .

[0086] The multi-level feature multicast strategy can be expressed as formula (25):

[0087] In formula (25), Representing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy can be expressed as formula (26):

[0088] In formula (26), represents the multicast bandwidth allocation strategy, and the power allocation strategy can be expressed as formula (27):

[0089] In formula (27), Denotes the power allocation strategy based on which the semantic loss and the average maximum multicast delay are minimized.

[0090] It should be noted that the steps of optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy to minimize the semantic loss and the average maximum multicast delay to obtain the optimal solution include: under any given multicast grouping strategy, jointly optimizing the corresponding bandwidth allocation strategy and the corresponding power allocation strategy to obtain the optimal bandwidth allocation strategy and the optimal power allocation strategy; based on the optimal bandwidth allocation strategy and the optimal power allocation strategy, obtaining the optimal multicast grouping strategy.

[0091] Exemplarily, the optimization problem can be modeled as P1, which can be expressed as formula (28) to formula (29).

[0092]

[0093] The P1 constraint condition is expressed as formula (30) to formula (34).

[0094]

[0095] The optimal multi-stage multicast selection strategy for problem (P1) is expressed as formula (36)

[0096] The bandwidth allocation strategy and power allocation strategy are expressed as formula (37):

[0097] The power allocation strategy is expressed as formula (38):

[0098] It should be noted that the meanings of the characters involved in formulas (28) to (38) have been specifically explained in the above formulas and will not be repeated here. The steps of obtaining the optimal multicast grouping strategy based on the optimal bandwidth allocation strategy and the optimal power allocation strategy include: generating all multicast grouping strategies based on the optimal bandwidth allocation strategy and the optimal power allocation strategy; searching for the minimum value of the semantic loss and the average maximum multicast delay among all multicast grouping strategies, and the multicast grouping strategy corresponding to the minimum value is the multicast grouping optimal strategy.

[0099] For example, the optimal solution to problem (P1) can be obtained based on brute force search. First, in any given multicast grouping strategy Under the bandwidth allocation strategy and power allocation strategy Then, the optimal grouping strategy is found through brute force search. In any given In this case, problem (P1) is equivalently transformed into problem (P2), and P2 can be expressed as formula (39).

[0100]

[0101] in, Indicates n The hierarchical multicast delay,P2 constraint is expressed as formula (40), formula (31), (32), (35), and formula (41).

[0102]

[0103] in, ， ， .

[0104] The algorithm for solving problem (P2) is summarized in Algorithm 1. In it, denotes the dual variable associated with constraint (40), β1 ≥ 0 and β2 ≥ 0 denotes the dual variables associated with constraint (31) and constraint (32), respectively.

[0105] It should be noted that given formula (42) to formula (45) are used in Algorithm 1.

[0106]

[0107] Among them, Algorithm 1 may include: initialization ,for , , ; According to the given , , , obtained from formulas (42), (43) and (44) , and ,in, Indicates that the input is , , Under the conditions nHierarchical multicast delay, Indicates that the input is , , Conditional allocation to l Layer multicast bandwidth, Indicates that the input is , , The distribution under the condition l Layer multicast power.

[0108] calculate about , , The subgradient of is given as follows:

[0109] Update accordingly according to the ellipse method , , , returns the value of , , , obtained from formulas (42), (43) and (44) ， and steps until , , Converge within the specified accuracy range; set up (g)= , obtained from equation (45) , and set formula (49)

[0110] In formula (49), It represents finding the optimal power allocation in a multi-user communication system while considering both power and bandwidth constraints.

[0111] It is understandable that by obtaining the optimal bandwidth allocation strategy and optimal power allocation strategy Then, the optimal strategy for multicast grouping is obtained through brute force search. , and finally the optimal solution to problem (P1) is obtained.

[0112] This embodiment determines the semantic losses corresponding to different levels according to the semantic errors of different levels, determines the multicast delays corresponding to different levels according to the communication resources of the data receiving devices in different levels, establishes an optimization problem according to the semantic losses corresponding to different levels and the multicast delays corresponding to different levels, and obtains the optimal solution. This process enables the entire MEC system to have efficient target recognition and classification performance, meets the needs of different mobile agents in information transmission, reduces transmission delay and resource consumption as a whole, and realizes efficient transmission of multi-level features. In one implementation, based on the first embodiment above, the specific multicast transmission steps may be as follows. For example, assuming that The image samples are sent to the mobile agent through the MEC server, denoted as . Each mobile agent needs all the image samples, so multicasting is considered at the MEC server. Based on the multi-level processing of the source information (picture information is taken as an example here) by the multi-level feature encoder, for each image sample to be transmitted, there are five transmission level types (features) to choose from at the MEC. By obtaining the multi-level features, the semantic error is obtained. Among them, the potential low-dimensional intermediate features can be obtained by extracting the visual features of the target image and projecting the visual features into the potential low-dimensional intermediate space. Based on the multi-level semantic encoder, the potential low-dimensional intermediate features are projected into the semantic feature space to obtain the semantic features. Based on the first semantic knowledge base (SKB at the MEC server), the semantic features are classified and judged to obtain the semantic error, and the semantic vector corresponding to the minimum semantic error can also be obtained. Among them, the first semantic knowledge base is a knowledge base maintained at the MEC, which is used to store the semantic vectors of each sample image. Further, the target feature is confirmed, and the MEC uses the multi-level semantic encoder to send the final target feature of the semantic information to the receiving end through the channel. In this process, the communication resources are optimized, the bandwidth is reasonably allocated, the bandwidth consumption is reduced, and the transmission delay is reduced. The receiving end decodes the semantic information features through a multi-level semantic decoder, matches it with its own semantic knowledge base, obtains the semantics to be transmitted, and completes the semantic communication process.

[0113] Among them, which feature is the target feature of the current target image can be determined according to the size relationship between the semantic error and the set error threshold, and at the same time, the interactive feedback action between the receiving end and the MEC server. For example, when the semantic error is less than the preset threshold, the estimated category information corresponding to the semantic error is used as the target feature of the current target image and sent to the data receiving end; the data receiving end feeds back the recognition result to the MEC. If the recognition result is that the estimated category semantic vector corresponding to the received estimated category identifier does not exist in the second semantic knowledge base (SKB of the mobile agent end) of the receiving end, then the MEC needs to resend the estimated category semantic vector as the target feature of the current target image to the data receiving end. By analogy, in the sending decision at each level, the corresponding threshold judgment will be made first. If it meets the threshold requirements, the current level will be sent; if it does not meet the requirements, it will retreat to the sending decision of the previous level and make a similar judgment again to finally determine the sending transmission strategy.

[0114] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the multi-level feature multicast transmission method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0115] This application also provides a multi-level feature multicast transmission device, please refer to Figure 5 , the multi-level feature multicast transmission device comprises: A feature acquisition module 10 is used to acquire multi-level features according to a target image; A transmission resource module 20, configured to determine semantic errors at different levels according to the multi-level features, and determine communication resources of a data receiving device at a current level according to target features of a target image; The transmission optimization module 30 is used to establish an optimization problem based on the semantic errors of different levels and the communication resources of data receiving devices of different levels to obtain an optimal solution; The multicast transmission module 40 is used to perform multicast transmission on the target feature based on the optimal solution.

[0116] The multi-level feature multicast transmission device provided by the present application adopts the multi-level feature multicast transmission method in the above embodiment, which can solve the technical problem of large communication resource overhead in multi-user scenarios. Compared with the prior art, the beneficial effects of the multi-level feature multicast transmission device provided by the present application are the same as the beneficial effects of the multi-level feature multicast transmission method provided by the above embodiment, and other technical features in the multi-level feature multicast transmission device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0117] The present application provides a multi-level feature multicast transmission device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the multi-level feature multicast transmission method in the above-mentioned embodiment one.

[0118] Reference below Figure 6 , which shows a schematic diagram of the structure of a multi-level feature multicast transmission device suitable for implementing the embodiment of the present application. The multi-level feature multicast transmission device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 6 The multi-level feature multicast transmission device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0119] like Figure 6 As shown, the multi-level feature multicast transmission device may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. Various programs and data required for the operation of the multi-level feature multicast transmission device are also stored in the RAM 1004. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the multi-level feature multicast transmission device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6A multi-layer feature multicast transmission device with various systems is shown, but it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have instead.

[0120] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0121] The multi-level feature multicast transmission device provided by the present application adopts the multi-level feature multicast transmission method in the above embodiment, which can solve the technical problem of large communication resource overhead in multi-user scenarios. Compared with the prior art, the beneficial effects of the multi-level feature multicast transmission device provided by the present application are the same as the beneficial effects of the multi-level feature multicast transmission method provided by the above embodiment, and other technical features in the multi-level feature multicast transmission device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0122] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0124] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, wherein the computer-readable program instructions are used to execute the multi-level feature multicast transmission method in the above-mentioned embodiment.

[0125] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0126] The above-mentioned computer-readable storage medium may be included in the multi-level feature multicast transmission device; or may exist independently without being assembled into the multi-level feature multicast transmission device.

[0127] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the multi-level feature multicast transmission device, the multi-level feature multicast transmission device enables the multi-level feature multicast transmission device to: obtain multi-level features according to the target image; determine the semantic errors of different levels according to the multi-level features, and determine the communication resources of the data receiving device in the current level according to the target features of the target image; establish an optimization problem based on the communication resources of the data receiving devices at different levels and the semantic errors at different levels to obtain the optimal solution; and multicast the target features based on the optimal solution.

[0128] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0129] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0130] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0131] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned multi-level feature multicast transmission method, and can solve the technical problem of high communication resource overhead in multi-user scenarios. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the multi-level feature multicast transmission method provided by the above-mentioned embodiment, and will not be repeated here.

[0132] The present application also provides a computer program product, including a computer program, which implements the steps of the multi-level feature multicast transmission method as described above when executed by a processor.

[0133] The computer program product provided by this application can solve the technical problem of high communication resource overhead in multi-user scenarios. Compared with the prior art, the beneficial effects of the computer program product provided by this application are the same as the beneficial effects of the multi-level feature multicast transmission method provided by the above embodiment, which will not be repeated here.

[0134] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A multi-level feature multicast transmission method, characterized in that: The method includes: Acquire multi-level features according to the target image, wherein the multi-level features include semantic features; Determining semantic errors at different levels according to the semantic features, and determining communication resources of a data receiving device at a current level according to target features of the target image; Based on the communication resources of data receiving devices at different levels and the semantic errors at the different levels, an optimization problem is established to obtain an optimal solution; The target features are multicast transmitted based on the optimal solution.

2. The method according to claim 1, characterized in that The step of establishing an optimization problem based on the semantic errors at different levels and the communication resources of the data receiving devices at different levels to obtain an optimal solution includes: Determining semantic losses corresponding to different levels according to the semantic errors of the different levels; Determining multicast delays corresponding to different levels according to communication resources of data receiving devices in the different levels; An optimization problem is established according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels to obtain an optimal solution.

3. The method according to claim 2, characterized in that The step of establishing an optimization problem according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels to obtain an optimal solution includes: Establishing a multi-level multicast selection strategy, a bandwidth allocation strategy, and a power allocation strategy according to the semantic losses corresponding to the different levels and the multicast delays corresponding to the different levels; By optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy, the semantic loss and the average maximum multicast delay are minimized to obtain an optimal solution.

4. The method according to claim 3, characterized in that The step of optimizing the multi-level feature multicast strategy, the multicast bandwidth allocation strategy, and the power allocation strategy to minimize semantic loss and average maximum multicast delay to obtain the optimal solution includes: Under any given multicast grouping strategy, the corresponding bandwidth allocation strategy and the corresponding power allocation strategy are jointly optimized to obtain the optimal bandwidth allocation strategy and the optimal power allocation strategy; Based on the optimal bandwidth allocation strategy and the optimal power allocation strategy, an optimal strategy for multicast grouping is obtained.

5. The method according to claim 4, characterized in that The step of obtaining the optimal strategy for multicast grouping based on the optimal bandwidth allocation strategy and the optimal power allocation strategy comprises: Based on the optimal bandwidth allocation strategy and the optimal power allocation strategy, generating all multicast grouping strategies; The minimum value of the semantic loss and the average maximum multicast delay is searched among all the multicast grouping strategies, and the multicast grouping strategy corresponding to the minimum value is the multicast grouping optimal strategy.

6. The method according to claim 1, characterized in that The multi-level features also include visual features and potential low-dimensional intermediate features. The step of obtaining the multi-level features according to the target image includes: Extracting visual features of the target image, and projecting the visual features into a potential low-dimensional intermediate space to obtain the potential low-dimensional intermediate features; Based on a multi-level semantic encoder, the potential low-dimensional intermediate features are projected into a semantic feature space to obtain the semantic features.

7. A multi-level feature multicast transmission device, characterized in that: The device comprises: A feature acquisition module, used to acquire multi-level features according to a target image, wherein the multi-level features include semantic features; A transmission resource module, used to determine the semantic errors of different levels according to the semantic features, and determine the communication resources of the data receiving device in the current level according to the target features of the target picture; A transmission optimization module, used to establish an optimization problem based on the semantic errors at different levels and the communication resources of data receiving devices at different levels to obtain an optimal solution; A multicast transmission module is used to perform multicast transmission on the target feature based on the optimal solution.

8. A multi-level feature multicast transmission device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the multi-level feature multicast transmission method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the multi-level feature multicast transmission method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the multi-level feature multicast transmission method according to any one of claims 1 to 6 are implemented.

Images