A pedestrian description generation method based on multi-model local description integration

Abstract

The present invention discloses a method for generating pedestrian descriptions based on the integration of multi-model local descriptions, and relates to the field of image data processing technology. It includes obtaining pedestrian images to be annotated, inputting the images and preset instructions into multiple image-text models, and generating local text descriptions of various attributes of the pedestrians; using an image-text matching model to extract image and text features, calculating the matching degree between the image and the text by cosine similarity, and screening out the attribute description that best matches the image from multiple descriptions; using a large language model to merge the screened local descriptions to generate a smooth text description of the pedestrian as a whole; training the image-text matching model through pedestrian image and text data sets, and optimizing it using contrast loss and matching loss functions until the overall loss function converges. The present invention generates attribute phrases for pedestrians through a large cross-modal model, making the descriptions of pedestrian attribute phrases richer and more diverse, and avoiding the convergence of generated descriptions.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method for generating pedestrian descriptions based on the integration of multi-model local descriptions. Background Art

[0002] The task of text-based person retrieval involves finding images of specific pedestrians that match a given text description from a large number of pedestrian images captured by cameras. In the field of text-based person retrieval, image data is relatively easy to obtain and can be automatically captured by surveillance cameras or other devices. However, detailed, high-quality pedestrian text descriptions typically require manual compilation or annotation, which is time-consuming and labor-intensive. With the development of large-scale pre-trained vision-language cross-modal models, these models, with their large parameter size and powerful computational power, can better understand complex scenes in images and generate more accurate text descriptions. However, due to the domain-specific and fine-grained nature of pedestrian text descriptions, direct application of these existing models to pedestrian image description tasks is not ideal. Currently, the most common approaches for generating pedestrian descriptions are based on predefined pedestrian attribute combinations and cross-modal large-scale models. The pre-defined pedestrian attribute combination generation method predefines a set of common pedestrian attributes, then uses a pre-trained model to select the most appropriate phrase from the set of pedestrian attributes. Finally, all attributes are combined to generate a pseudo-text description.

[0003] However, while predefined pedestrian attributes cover common pedestrian appearance characteristics, they cannot fully express individual differences in more complex and diverse pedestrian scenes, making it difficult to obtain diverse pedestrian descriptions. Generative methods based on large cross-modal models generate pedestrian descriptions through brief instructions. Generally, a single round of dialogue can generate a description of an image. However, due to the inherent ambiguity of pedestrian images, some generated descriptions may be irrelevant to the objective content of the image.

[0004] Currently, the common pedestrian description generation methods can be summarized as: generation methods based on predefined pedestrian attribute combinations and generation methods based on cross-modal large models, but there are still many problems. For example: in the generation method based on predefined pedestrian attribute combinations, since the attributes of pedestrians in actual scenarios are changeable, the predefined attribute combinations cannot cover all potential pedestrian characteristics, and the generated descriptions are easily restricted by predefined templates, lacking diversity and accuracy in detail descriptions; when the cross-modal large model is used to directly generate the overall description of the pedestrian image, it is impossible to generate all the pedestrian attributes in the image at one time, and some more common pedestrian attribute descriptions are often omitted. At the same time, the background information in the image may interfere with the model, causing the generated description to contain content that is not related to the pedestrian. Summary of the Invention

[0005] In view of the problems existing in the above background technology, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to generate attribute phrases of pedestrians through a large cross-modal model, describe the attributes of pedestrians through multiple graph-to-text models, screen the generated results, eliminate possible erroneous descriptions, and improve the accuracy of text descriptions.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a method for generating pedestrian descriptions based on the integration of multi-model local descriptions, which includes obtaining a pedestrian image to be annotated, inputting the image and preset instructions into multiple image-text models, and generating local text descriptions of various attributes of the pedestrian; using an image-text matching model to extract image and text features, calculating the matching degree between the image and the text by cosine similarity, and screening out the attribute description that best matches the image from multiple descriptions; using a large language model to merge the screened local descriptions to generate a fluent text description of the pedestrian as a whole; training the image-text matching model using pedestrian image and text datasets, and optimizing it using contrast loss and matching loss functions until the overall loss function converges.

[0009] As a preferred solution of the pedestrian description generation method based on the integration of multi-model local descriptions described in the present invention, the method of using a large language model to merge the screened local descriptions to generate a fluent text description of the pedestrian as a whole includes the following steps: performing a fluent text description of the pedestrian as a whole; extracting attribute phrases in the pedestrian text, combining bidirectional attention for weighting, and constructing a masked phrase modeling loss function.

[0010] As a preferred solution of the pedestrian description generation method based on the integration of multi-model local descriptions of the present invention, the fluent text description of the pedestrian as a whole includes:

[0011] ,

[0012] = ,

[0013] ,

[0014] Among them, LLM is a large language model, is the final merged text description, , ,..., Represents the attributes of pedestrian images, is the i-th cross-modal large model, is the instruction hint designed for the jth attribute, I is the image, K is the number of attributes in the pedestrian image, For the attributes after filtering A local text description of .

[0015] As a preferred solution of the pedestrian description generation method based on multi-model local description integration of the present invention, the construction of the mask phrase modeling loss function includes the following steps: And the image representation F I As the input of the cross-modal encoder, the internal calculation of the cross-attention layer is performed; according to the bidirectional attention mechanism, the features of each image block are weighted to obtain the weighted global representation; the image is calculated and phrases The coarse similarity between them is measured; bidirectional attention is used to weight local attributes for local alignment; and a mask phrase modeling loss function is designed.

[0016] As a preferred solution of the pedestrian description generation method based on multi-model local description integration of the present invention, the cross-attention layer internal calculation includes:

[0017] ,

[0018] ,

[0019] in, and H is the number of heads, , , is the linear projection matrix, yes The transpose of ,A attention relation matrix describes the similarity relationship between phrase tokens and image patches.

[0020] As a preferred solution of the pedestrian description generation method based on multi-model local description integration of the present invention, wherein: the weighting of the features of each image block according to the bidirectional attention mechanism includes:

[0021] ,

[0022] in reflects the weight of the j-th image block to the input phrase, is the j-th image block feature of the input, is the number of image patches in the input image.

[0023] As a preferred solution of the pedestrian description generation method based on multi-model local description integration of the present invention, the training loss of the image-text matching model is:

[0024] ,

[0025] in, is the set of phrases in the input text.

[0026] In the second aspect, an embodiment of the present invention provides a pedestrian description generation system based on the integration of multi-model local descriptions, which includes: an image acquisition module for acquiring pedestrian images to be annotated; a description generation module for inputting images and preset instructions into multiple image-text models to generate local text descriptions of various attributes of pedestrians; a feature extraction and matching module for extracting image and text features using an image-text matching model, calculating the matching degree between the image and the text through cosine similarity, and screening out the attribute description that best matches the image from multiple descriptions; a description integration module for merging the screened local descriptions using a large language model to generate a fluent text description of the pedestrian as a whole; a model training module for training the image-text matching model using pedestrian image and text datasets, and optimizing the model using contrast loss and matching loss functions until the overall loss function converges.

[0027] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the pedestrian description generation method based on the integration of multi-model local descriptions as described in the first aspect of the present invention are implemented.

[0028] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the pedestrian description generation method based on the integration of multi-model local descriptions as described in the first aspect of the present invention are implemented.

[0029] The beneficial effects of the present invention are as follows: the present invention generates attribute phrases of pedestrians through a cross-modal large model, making the descriptions of pedestrian attribute phrases richer and more diverse, avoiding the convergence of generated descriptions and the inability to reflect the detailed differences between pedestrians; compared with the traditional generation method based on a cross-modal large model, the present invention describes pedestrian attributes through multiple graph-to-text models, and screens the generated results to eliminate possible erroneous descriptions, thereby improving the accuracy of the text description. The present invention provides a solution to the problem of the diversity of pedestrian description generation in the past and the difficulty of ensuring the accuracy of global descriptions generated by a single model. It enriches the pedestrian attributes in the generated text, eliminates some text descriptions that are not related to the pedestrians themselves, and improves the accuracy of the text description. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0031] Figure 1 This is a process for generating pedestrian descriptions based on the integration of multi-model local descriptions.

[0032] Figure 2 Schematic diagram of the training process of the pedestrian image-text matching model.

[0033] Figure 3 Schematic diagram of the global loss of training for the image-text matching model.

[0034] Figure 4 Schematic diagram of the local loss of training for the image-text matching model.

[0035] Figure 5 Generate a structure diagram for a local description.

[0036] Figure 6 is the fusion triplet loss hyperparameter Graph of experimental data. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0040] Example 1

[0041] Reference Figures 1 to 6 , which is the first embodiment of the present invention, provides a pedestrian description generation method based on the integration of multi-model local descriptions, including:

[0042] S1: Obtain the pedestrian image to be annotated, input the image and preset instructions into multiple image-to-text models, and generate local text descriptions of various attributes of the pedestrian.

[0043] S1.1: Obtain the pedestrian image to be annotated.

[0044] S1.2: Input the pedestrian image and pre-designed instruction prompts into n image-to-text models, and obtain n texts for K attributes:

[0045] ,

[0046] ,

[0047] in, is the i-th cross-modal large model, is the instruction hint designed for the jth attribute, I is the pedestrian image, , ,..., Represents attributes in pedestrian images, which can be different aspects of pedestrian appearance or behavior characteristics, such as gender, top style and color, etc.

[0048] S2: Use the image-text matching model to extract image and text features, calculate the matching degree between the image and text through cosine similarity, and filter out the attribute description that best matches the image from multiple descriptions.

[0049] S2.1: Build a pedestrian image-text matching model.

[0050] Preferably, Figure 2 The figure shows the training process of the pedestrian image-text matching model. The constructed pedestrian image-text matching model structure is as follows: Figure 3 shown.

[0051] Specifically, the architecture is based on the transformer model, with image and text encoders, corresponding momentum models and a cross-modal encoder. Specifically, ViT is used as the image encoder, the first 6 layers of BERT are used as the text encoder, and the last 6 layers are used as the cross-modal encoder.

[0052] S2.2: Obtain a pedestrian image and text dataset, and extract image and text features for each pair of pedestrian image and text.

[0053] Construct image-text loss function, image-text matching loss function and triplet loss function.

[0054] Here, the image-text contrast loss is used ( ) and image-text matching loss ( ) for global alignment and fuses the triplet loss ( ) to further enhance the global alignment and thus enhance the difference between positive and negative text-image pairs.

[0055] The specific calculation steps are as follows:

[0056] Image-text pair The image features obtained by the image input image encoder and the text , text input text encoder to get text features :

[0057] ,

[0058] ,

[0059] Among them, d represents the feature dimension of the image or text after encoding , is the number of image patches in the input image and is the number of tokens in the input text.

[0060] Specifically, the cosine similarity between text embedding and image embedding is calculated to obtain their rough similarity, which can be used to calculate the rough similarity between image and text:

[0061] ,

[0062] in, represents the norm, (or ) is to (or ) is a linear projection onto a representation with lower dimensionality, (or ) represents the global representation of text (or image).

[0063] During training, the latest images and text from the kinetic encoder The global representation is stored in two queues and In , image-to-text and text-to-image similarities are computed by combining representations from the original encoder and the paired momentum encoder as follows:

[0064] ,

[0065] ,

[0066] in, Measure the global representation of the image and the A global representation of a text (in The rough similarity between , is a learnable temperature parameter, exp( ) represents an exponential operation with a constant e as the base.

[0067] Therefore, the image-text contrast loss Defined as:

[0068] ,

[0069] in, , ∈ It represents the label relationship between image and text, with the value of image-text pair being 1 and the value of negative pair being 0. Represents a function for calculating the cross entropy loss.

[0070] This constructs a contrast loss between image and text, which increases the ability of the pedestrian image-text matching model to distinguish different pairs.

[0071] Afterwards, and Input the cross-modal encoder to obtain the image-text fusion representation:

[0072] ;

[0073] On this basis, using the linear projection matrix W O Compute the ground truth similarity between an image and text:

[0074] ,

[0075] in Represents the similarity between the global image representation and text after fusion with the cross-modal encoder.

[0076] Based on the fine-grained similarity, the image-text matching loss can be expressed as:

[0077] ,

[0078] in, Indicates whether the image-text pair is a positive number pair, 1 indicates a positive number, and 0 indicates a negative number.

[0079] Typically, negative examples are sampled from the training batch.

[0080] Although the image-text matching loss is effective in promoting the training of cross-modal encoders, it often prioritizes individual pairs without considering the contrast between pairs.

[0081] Specifically, with negative and( , T) compared to Should output higher ,in in represents text that does not match image I, Same thing.

[0082] To satisfy it, in addition to the baseline and In addition, a fusion triplet loss is added to emphasize the gap between positive and negative samples:

[0083] ,

[0084] in, is the minimum margin between positive and negative pairs, represents max(·,0), and experiments show that Taking 0.6 is the best effect, the specific experimental data are as follows Figure 6 As shown, the R@k metric represents the probability that at least one matching person's image appears in the top k candidate lists given a query description. It can be seen that here The best effect is achieved when the value is 0.6.

[0085] S3: Use a large language model to merge the filtered local descriptions to generate a fluent text description of the pedestrian as a whole.

[0086] S3.1: Preferably, a fluent textual description of the pedestrian as a whole is as follows:

[0087] ,

[0088] Among them, LLM is a large language model, A textual description of the final merge.

[0089] Attribute phrases are extracted from pedestrian texts, weighted by bidirectional attention, and a masked phrase modeling loss function is constructed.

[0090] S3.2: Extract attribute phrases from pedestrian text, combine bidirectional attention for weighting, and construct a masked phrase modeling loss function.

[0091] Here, in order to improve the ability of local attribute matching, the NLTK natural language processing library is used to extract attribute phrases, and the gradient is used as backward attention weighting to assist in the local alignment of pedestrian images and local pedestrian attribute phrases.

[0092] The phrase represents And the image representation F IAs the input of the cross-modal encoder, the calculation process inside the cross-attention layer is:

[0093] ,

[0094] ,

[0095] in, and H is the number of heads, , , is the linear projection matrix, Q, K, and V represent the query vector, key vector, and value vector in the Transformer architecture, respectively. yes The transpose of ,A attention relation matrix describes the similarity relationship between phrase tokens and image patches.

[0096] The forward attention of the representation of the jth image patch on the global features of the phrase is:

[0097] ,

[0098] in, is the first row and jth column of A, Fa is the flag of forward attention, represents the forward attention weight of the j-th image patch to the phrase.

[0099] The predicted score of the masked word label is:

[0100] ,

[0101] in, is a linear projection, represents the relationship of each image patch to the masked token in the phrase of A.

[0102] The backward attention based on the back-propagated gradient of s is expressed as:

[0103] ,

[0104] in, Represents an image block The backward attention weights for the predicted scores s of the masked word labels are obtained by using the backpropagation gradient To calculate, is an image block No. dimensional eigenvalues, Represents a linear projection matrix No. elements.

[0105] Finally, bidirectional attention is defined as:

[0106] ,

[0107] in, and They are and The average value on the attention head, norm is a function that normalizes the weights.

[0108] According to the bidirectional attention mechanism, the features of each image block are weighted to obtain the weighted global representation:

[0109] ,

[0110] in reflects the weight of the j-th image block to the input phrase, is the number of image patches in the input image.

[0111] therefore, , which is the sum of the representations of all weighted image patches, can also be regarded as a phrase-guided global image representation. The representation of the image patch corresponding to the phrase is highlighted, while the representations of other patches are weakened.

[0112] Therefore, the image can be calculated and phrases The rough similarity between:

[0113] ,

[0114] in, The norm of , (resp. ) is to (or ) is mapped to a linear projection of a representation with lower dimensionality.

[0115] In summary, the bidirectional attention-weighted local alignment loss for an image-phrase pair is:

[0116] ,

[0117] In addition to the bidirectional attention-weighted local alignment for local attributes, a masked phrase modeling loss function is proposed.

[0118] Specifically, in the input phrase A masking operation is performed on the phrase by randomly replacing one word in the phrase with the token [MASK]. The masked phrase is represented as ,Will Input into the image-text matching model and calculate the final fusion representation:

[0119] ,

[0120] in Enter the phrase The number of tokens after passing through the text encoder, represents the text representation of the masked phrase, where represents the global representation of the masked phrase, and Then it represents the local representation of each text token.

[0121] Then, a multi-layer perceptron (MLP) classifier is used to predict the original tokens in the phrase using an MPM loss on the image-phrase pair:

[0122] ,

[0123] in, is the one-hot vocabulary distribution of the j-th token, is the size of the vocabulary.

[0124] S4: Train the image-text matching model using pedestrian image and text datasets, and optimize the model using contrast loss and matching loss functions until the overall loss function converges.

[0125] Preferably, the training loss of the entire model is:

[0126] ,

[0127] in, is the set of phrases in the input text.

[0128] During the specific training, AdamW was used as the optimizer and the initial learning rate was set to 1×10 -5 And the learning rate is adjusted using the cosine learning rate decay method.

[0129] Furthermore, this embodiment also provides a pedestrian description generation system based on the integration of multi-model local descriptions, including:

[0130] The image acquisition module is used to obtain images of pedestrians to be annotated; the description generation module is used to input images and preset instructions into multiple image-text models to generate local text descriptions of various attributes of pedestrians; the feature extraction and matching module is used to extract image and text features using the image-text matching model, calculate the matching degree between the image and text through cosine similarity, and screen out the attribute description that best matches the image from multiple descriptions; the description integration module is used to use a large language model to merge the screened local descriptions to generate a fluent text description of the pedestrian as a whole; the model training module is used to train the image-text matching model using pedestrian image and text datasets, and optimize the model using contrast loss and matching loss functions until the overall loss function converges.

[0131] This embodiment also provides a computer device, which is suitable for the pedestrian description generation method based on the integration of multi-model local descriptions, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the pedestrian description generation method based on the integration of multi-model local descriptions proposed in the above embodiment.

[0132] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0133] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the pedestrian description generation method based on the integration of multi-model local descriptions as proposed in the above embodiment.

[0134] In summary, the present invention generates attribute phrases for pedestrians through a cross-modal large model, making the descriptions of pedestrian attribute phrases richer and more diverse, avoiding the convergence of generated descriptions and the inability to reflect the detailed differences between pedestrians; compared to traditional generation methods based on cross-modal large models, the present invention describes pedestrian attributes through multiple graph-to-text models, and screens the generated results to eliminate possible erroneous descriptions, thereby improving the accuracy of the text description. The present invention is a solution to the problem of the diversity of pedestrian description generation in the past and the difficulty of ensuring the accuracy of global descriptions generated by a single model. It enriches the pedestrian attributes in the generated text, eliminates some text descriptions that are not related to the pedestrians themselves, and improves the accuracy of the text description.

[0135] Example 2

[0136] The second embodiment of the present invention provides a pedestrian description generation method based on the integration of multi-model local descriptions. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0137] Taking the unlabeled pedestrian dataset LUPerson as an example, the steps are as follows:

[0138] The unlabeled dataset LUPerson is selected, which contains a large number of unlabeled pedestrian images;

[0139] The pedestrian image I and the instruction prompts designed for the K attributes of the pedestrian are input into n cross-modal large models respectively. Specifically, the pedestrian image I and the instruction prompts for the jth (1 j K) command prompts designed for each attribute Input to the i-th (1 i n) Cross-modal large models In the example, we get the text description of the jth attribute of the pedestrian image I by the model , then for the jth attribute, n different text descriptions can be generated ;

[0140] Given a pair of input images and their associated attribute text , the input image is processed by the image encoder Process it to extract its feature representation and get the image representation ,in, represents the global representation of the image, and It represents the representation of the image block. The text encoder will target the attribute Each text Encode and generate text representation ,in, represents the global representation of the text, and The corresponding text mark;

[0141] Subsequently, the similarity between the image and text embeddings is estimated as in and is a linear transformation used to embed the global and Mapped to a standardized low-dimensional representation, the similarity between the text features and image features generated by each model is calculated to obtain , by calculating and comparing the cosine similarity value Sort in descending order to define the overall match score, then use the sorted index argsort( ) for the generated n pairs The text description of the attribute is sorted. According to this sort, the attribute description text that is found last is selected as the attribute The final description is recorded as = .

[0142] Select the attribute description text Combine the attribute name A with the corresponding attribute description text T A Combinations and prompt words The large language model (LLM) generates the most likely text description by maximizing the conditional probability. In this process, the model predicts the probability distribution of the next word word by word ,in Is generated text The words, is generated before Combine multiple local attribute description texts into a smooth and complete pedestrian text description.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A pedestrian description generation method based on the integration of multi-model local descriptions, characterized by: include, Obtain an image of a pedestrian to be annotated, input the image and preset instructions into multiple image-to-text models, and generate local text descriptions of various attributes of the pedestrian; Extracting image features and text features of each local text description using an image-text matching model, calculating the matching degree between the image features and the text features using cosine similarity, and selecting the local text description that best matches the image from multiple local text descriptions; A large language model is used to merge the filtered local text descriptions to generate a smooth text description of the pedestrian as a whole; The overall fluent text description of the pedestrian includes: Among them, LLM is the large language model, T is the final merged text description, A1, A2, ..., A K Represents the attributes of pedestrian images, MLLM i is the i-th cross-modal large model, is the instruction hint designed for the jth attribute, I is the image, K is the number of attributes in the pedestrian image; is the filtered local text description of attribute A1; is the filtered local text description of attribute A2; For attribute A after filtering K A local text description of For attribute A after filtering j A local text description of j Represents the jth attribute in the pedestrian image; π1 indicates that after similarity calculation, the similarity is ranked in descending order as π1, indicating the highest similarity; For attribute A j After similarity calculation, the local text description with the highest similarity is obtained after descending sorting; For attribute A j The attributed text generated in the i-th cross-modal model.

2. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the pedestrian description generation method based on multi-model local description integration according to claim 1 are implemented.

3. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for generating pedestrian descriptions based on the integration of multi-model local descriptions according to claim 1 are implemented.

Images