CN116385600A - Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment - Google Patents
Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment Download PDFInfo
- Publication number
- CN116385600A CN116385600A CN202310376114.5A CN202310376114A CN116385600A CN 116385600 A CN116385600 A CN 116385600A CN 202310376114 A CN202310376114 A CN 202310376114A CN 116385600 A CN116385600 A CN 116385600A
- Authority
- CN
- China
- Prior art keywords
- target
- remote sensing
- sensing image
- word
- training
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012512 characterization method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000012216 screening Methods 0.000 claims abstract description 26
- 230000014509 gene expression Effects 0.000 claims abstract description 25
- 238000000034 method Methods 0.000 claims abstract description 22
- 239000013598 vector Substances 0.000 claims description 52
- 230000011218 segmentation Effects 0.000 claims description 18
- 238000002372 labelling Methods 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 238000011160 research Methods 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/60—Editing figures and text; Combining figures or text
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9532—Query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The invention relates to a distributed characterization method and device for target characteristics of a remote sensing image and electronic equipment, wherein the method comprises the following steps: expressing the target characteristics by using prior information of unstructured text; screening unstructured text expression by combining the structured remote sensing image knowledge; training a semantic environment model of the remote sensing image, and representing target characteristics in the remote sensing image in a distributed mode. By implementing the scheme of the invention, the target feature expression combining unstructured text knowledge and structured image knowledge can be realized.
Description
Technical Field
The invention relates to the technical field of target characteristic analysis, in particular to a remote sensing image target characteristic distributed characterization method, system and electronic equipment based on a neuro-language model.
Background
Because of complex multi-source aircraft target characteristics and the correlation among information, in the problem of remote sensing image aircraft target characteristic expression, how to apply experience information in unstructured text data to structured image characteristic expression and to simplify and effectively express multi-source discrete aircraft target characteristics becomes a difficulty. The current mainstream unstructured text analysis method is to perform distributed word vector expression by a natural language processing method and extract feature information contained in a text. For the extraction of the structural image data characteristics, a mode of manually screening and manufacturing target characteristic samples is mainly relied on, and the mode is low in efficiency and difficult to obtain a good effect. Therefore, unstructured target features obtained through natural language processing are used as priori, the unstructured target features are applied to the characteristic characterization process of the airplane target features of the structured remote sensing images, and the improvement of semantic expression efficiency of the airplane target features becomes a research key point.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention aims to provide a distributed characterization method, a distributed characterization system and electronic equipment for remote sensing image target characteristics based on a neuro-language model, which can realize target characteristic expression combining unstructured text knowledge and structured image knowledge.
In order to achieve the above purpose, the technical scheme of the invention is as follows:
a first aspect of an embodiment of the present invention provides a distributed characterization method of a target characteristic of a remote sensing image, including:
expressing the target characteristics by using prior information of unstructured text;
screening unstructured text expression by combining the structured remote sensing image knowledge;
training a semantic environment model of the remote sensing image, and representing target characteristics in the remote sensing image in a distributed mode.
According to a first aspect of the embodiment of the present invention, the expressing the target characteristic by using the prior information of the unstructured text includes:
acquiring sufficient text data, and constructing a text data set with high correlation degree with the target through screening and preprocessing;
word segmentation is carried out on the text data set by using a word segmentation tool, and word frequency of words is counted;
training and learning the segmented data set by using the word vector model to obtain the word vector corresponding to each segmented word.
According to a first aspect of an embodiment of the present invention, the obtaining sufficient text data, and constructing a text data set with high correlation with the target through screening and preprocessing, includes: and obtaining a large amount of documents by using the subject keywords related to the target through a search engine, screening out documents which do not accord with the research subject, eliminating messy codes, and establishing a text data set with high degree of relevance to the target.
According to a first aspect of the embodiment of the present invention, the training and learning the segmented data set by using the word vector model includes:
dividing the segmented data set into a word segmentation training set and a word segmentation testing set, initializing a pre-training model by using a sliding window neural network skip-gram and taking word2vec as a basis, and training the word segmentation training set to obtain a word vector model which is suitable for the target field;
characterizing each word in the word segmentation test set by utilizing the word vector model to obtain word vectors, fusing the word vectors and extracting keywords from the word vectors;
and characterizing the keywords by using the word vector model again to obtain word vectors corresponding to each keyword.
According to the first aspect of the embodiment of the invention, when the unstructured text expression is screened by combining the knowledge of the structured remote sensing image, the screening basis is whether keywords obtained by the unstructured text expression according to the target characteristics can be marked and represented in the structured remote sensing image.
According to a first aspect of the embodiment of the present invention, the training of the semantic environmental model of the remote sensing image, and the distributed characterization of the target characteristics in the remote sensing image, includes:
randomly acquiring a structured remote sensing image set, and labeling a target represented in the structured remote sensing image set by a keyword obtained by unstructured text expression of the target characteristic to obtain a target characteristic vector;
superposing the target characteristic vector and the word vector corresponding to the keyword, and training the neuro-language model to obtain a remote sensing image semantic environment model;
and carrying out distributed characterization on the target characteristics in the original structured remote sensing image set by using the remote sensing image semantic environment model.
According to a first aspect of the embodiment of the present invention, the labeling content of the target represented in the structured remote sensing image set by the keyword obtained by unstructured text expression of the target characteristic includes: the time and space attributes of the target specifically comprise a target name, a target ID, image shooting time, a longitude and latitude position of a target center, four corner points of a minimum circumscribed rectangular frame of the target, a target slice size, spatial resolution and background, and are represented by vectors.
A second aspect of the embodiments of the present invention provides a distributed characterization system for a target characteristic of a remote sensing image, implemented by using a distributed characterization method for the target characteristic of the remote sensing image, including:
the text characterization module is used for expressing the target characteristics by using prior information of unstructured text;
the screening module is used for screening unstructured text expression by combining the knowledge of the structured remote sensing image;
the distributed characterization module is used for training a semantic environment model of the remote sensing image and characterizing target characteristics in the remote sensing image in a distributed manner.
A third aspect of the embodiments of the present invention provides an electronic device comprising a memory and a processor, the memory having stored therein a computer program which, when executed by the processor, causes the processor to implement a method as described above.
A fourth aspect of embodiments of the invention provides a computer readable storage medium having stored thereon a computer program which when executed by a processor implements a method as described above.
Compared with the prior art, the invention has the following beneficial effects:
according to the scheme provided by the embodiment of the invention, the characteristic suitable for labeling of the structural remote sensing image is screened by combining the prior information of the airplane target characteristic word vector after the unstructured text characteristic extraction processing, the characteristic information is labeled by the superimposed image, and the distributed characterization of the airplane characteristic in the structural remote sensing image is completed by means of the deep-learning remote sensing image semantic environment model. Compared with the traditional method, semantic information and association relation of the target characteristics in the structured text and the unstructured image can be deeply mined, and the subsequent application effect based on the target characteristics of the airplane can be improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are required to be used in the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained from these drawings without inventive effort for a person of ordinary skill in the art.
Fig. 1 shows a schematic implementation flow diagram of a distributed characterization method of remote sensing image aircraft target characteristics based on a neuro-linguistic model according to an embodiment of the present invention.
Detailed Description
The description of the embodiments of this specification should be taken in conjunction with the accompanying drawings, which are a complete description of the embodiments. In the drawings, the shape or thickness of the embodiments may be enlarged and indicated simply or conveniently. Furthermore, portions of the structures in the drawings will be described in terms of separate descriptions, and it should be noted that elements not shown or described in the drawings are in a form known to those of ordinary skill in the art.
Any references to directions and orientations in the description of the embodiments herein are for convenience only and should not be construed as limiting the scope of the invention in any way. The following description of the preferred embodiments will refer to combinations of features, which may be present alone or in combination, and the invention is not particularly limited to the preferred embodiments. The scope of the invention is defined by the claims.
As shown in fig. 1, a first aspect of the embodiment of the present invention provides a distributed characterization method for remote sensing image aircraft target characteristics based on a neuro-language model, which specifically includes the following steps:
s110, expressing target characteristics by using prior information of unstructured text;
s120, screening unstructured text expression by combining the structured remote sensing image knowledge;
s130, training a semantic environment model of the remote sensing image, and representing target characteristics in the remote sensing image in a distributed mode.
In an exemplary embodiment, the specific implementation process of expressing the target feature by using the a priori information of the unstructured text in the step S110 includes the following steps:
first, sufficient text data is acquired, and a text data set with high correlation with a target is constructed through screening and preprocessing. The method specifically comprises the following steps: and obtaining a large amount of documents by using the subject keywords related to the target through a search engine, screening out documents which do not accord with the research subject, eliminating messy codes, and establishing a text data set with high degree of relevance to the target. For example, through advanced searching in the knowledge network by using topic keywords such as "airplane", "feature", 12885 related documents are obtained, and M text data sets with high correlation degree with the airplane field are established through preprocessing operations such as screening and rejecting.
Then, the word segmentation tool is used for segmenting the text data set, and word frequency of the vocabulary is counted. The word segmentation tool may be jieba, etc. The word segmentation data set is constructed by screening keywords according to the occurrence frequency of the word segmentation vocabulary in the acquired documents, so that the number of the keywords can be controlled, and keywords with higher correlation degree with the studied target subject field can be screened.
And training and learning the segmented data set by using a word vector model to obtain a word vector corresponding to each segmented word.
Specifically, the training and learning process for the segmented data set by using the word vector model includes the following steps:
dividing a segmented data set (segmented data set) into a segmented training set and a segmented test set, and initializing a pre-training model by using a sliding window neural network skip-gram and taking word2vec as a basis to train the segmented training set to obtain a word vector model which is suitable for a target field (such as an airplane field);
characterizing each word in the word segmentation test set by using a word vector model to obtain word vectors, fusing the word vectors, and extracting keywords W from the word vectors;
representing the keywords by using a word vector model again to obtain N-dimensional word vectors corresponding to each keyword
In an exemplary embodiment, when the unstructured text expression is screened in combination with the knowledge of the structured remote sensing image in step S120, the screening is based on whether the keyword obtained by the unstructured text expression is the target feature can be represented by labeling in the structured remote sensing image. If the representation can be marked, reserving the keyword, and taking the word vector corresponding to the keyword as an initial input vector of a language environment model of the subsequent remote sensing image.
In an exemplary embodiment, the training of the semantic environmental model of the remote sensing image in the step S130, and the implementation process of the distributed characterization of the target characteristics in the remote sensing image, include the following steps:
randomly acquiring a structured remote sensing image set, and labeling a target represented in the structured remote sensing image set by a keyword obtained by the target characteristics through unstructured text expression to obtain a target characteristic vector.
Specifically, the labeling of the target represented in the structured remote sensing image set by the keyword (for example 737, airport, etc.) obtained by unstructured text expression for the target characteristics includes: the time and space attributes of the target include the target name, target ID, image shooting time, longitude and latitude position of the target center, and minimum circumscribed rectangle frame I of the target Target object Is (x) 1 ,y 1 ),(x 2 ,y 2 ),(x 3 ,y 3 ),(x 4 ,y 4 ) Target slice size (length x width), spatial resolution (W x H), background, etc., and using N ′ The dimension vector representation may be
To vector the target characteristicWord vector corresponding to keyword ++> Superposition is performed, and the N "=n' +w-dimension vector +_n obtained after superposition is used>Training neuro-linguistic modelAnd obtaining a semantic environment model of the remote sensing image. The neural language model is a convolutional neural network CNN.
And carrying out distributed characterization on the target characteristics in the original structured remote sensing image set by using the trained remote sensing image semantic environment model. That is, distributed features of various targets that contain unstructured and structured information may be obtained. The target characteristic distributed feature obtained here can be used as a parameter to further participate in training of other tasks or be used alone. For example, aircraft targets may be classified by computing vector distances between distributed representations of aircraft target features, resulting in feature clustering topics.
Wherein word2vec model is used for aiming at target keyword w according to context i And (3) predicting:
p(w i |context)=p(w i |w i-k ,w i-k+1 ,…,w i-1 ,w i+1 ,…,w i+k-1 ,w i+k )
the cost function used by the word2vec model is as follows:
wherein k represents the sliding window size of the neural network skip-gram, and T represents the size of the word segmentation dataset used for training; p (w) in skip-gram model i+j |w i ) Defined by the Softmax function:
wherein w is I And w O Representing the input and output of the keyword w respectively,and->Input word vector and output word representing keyword w, respectivelyVector (interrelating keywords), V represents dictionary length.
The second aspect of the embodiment of the invention discloses a distributed characterization system of remote sensing image target characteristics, which is realized by using the distributed characterization method of remote sensing image target characteristics, and can be applied to various computers, notebook computers and other electronic equipment. The distributed characterization system for the target characteristics of the remote sensing image specifically comprises the following steps: the system comprises a text characterization module, a screening module and a distributed characterization module. The text characterization module is used for expressing the target characteristics by using prior information of unstructured text; the screening module is used for screening unstructured text expression by combining the knowledge of the structured remote sensing image; the distributed characterization module is used for training a semantic environment model of the remote sensing image and characterizing target characteristics in the remote sensing image in a distributed manner.
A third aspect of embodiments of the invention discloses an electronic device that may include a memory and a processor. The memory stores a computer program with executable program codes, and the processor calls the computer program stored in the memory to execute the distributed characterization method of the remote sensing image target characteristics disclosed by the embodiment of the invention.
A fourth aspect of the embodiments of the present invention discloses a computer readable storage medium, on which a computer program is stored, which when executed by a processor, causes the processor to implement the above-described distributed characterization method for remote sensing image target characteristics disclosed in the embodiments of the present invention.
According to the scheme disclosed by the embodiment of the invention, when the aircraft target feature in the remote sensing image is characterized, the prior information in the existing unstructured text data is fully referenced, and compared with the traditional experience summarization mode, the word2vec model is used for training the documents in the related field to obtain the aircraft target feature word vector with stronger pertinence, so that the efficient positioning summarization of the research target information is realized. On the basis, aircraft target features capable of being subjected to image annotation are screened, a remote sensing image semantic environment model is constructed, aircraft target features in a structural image are characterized, text feature word vectors and image annotation information vectors are combined, the remote sensing image semantic environment model is obtained through training, distributed representation of aircraft target features is achieved, and overall representation of the aircraft target features is enhanced.
The sequence numbers of the steps related to the method of the present invention do not mean the sequence of the execution sequence of the method, and the execution sequence of the steps should be determined by the functions and the internal logic, and should not limit the implementation process of the embodiment of the present invention in any way.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather to enable any modification, equivalent replacement, improvement or the like to be made within the spirit and principles of the invention.
Claims (10)
1. A distributed characterization method of remote sensing image target characteristics, comprising:
expressing the target characteristics by using prior information of unstructured text;
screening unstructured text expression by combining the structured remote sensing image knowledge;
training a semantic environment model of the remote sensing image, and representing target characteristics in the remote sensing image in a distributed mode.
2. The method of claim 1, wherein expressing the target property using a priori information of the unstructured text comprises:
acquiring sufficient text data, and constructing a text data set with high correlation degree with the target through screening and preprocessing;
word segmentation is carried out on the text data set by using a word segmentation tool, and word frequency of words is counted;
training and learning the segmented data set by using the word vector model to obtain the word vector corresponding to each segmented word.
3. The method of claim 2, wherein said obtaining sufficient text data to construct a text dataset with high relevance to said target by screening and preprocessing comprises: and obtaining a large amount of documents by using the subject keywords related to the target through a search engine, screening out documents which do not accord with the research subject, eliminating messy codes, and establishing a text data set with high degree of relevance to the target.
4. The method of claim 2, wherein training and learning the segmented data set using a word vector model comprises:
dividing the segmented data set into a word segmentation training set and a word segmentation testing set, initializing a pre-training model by using a sliding window neural network skip-gram and taking word2vec as a basis, and training the word segmentation training set to obtain a word vector model which is suitable for the target field;
characterizing each word in the word segmentation test set by utilizing the word vector model to obtain word vectors, fusing the word vectors and extracting keywords from the word vectors;
and characterizing the keywords by using the word vector model again to obtain word vectors corresponding to each keyword.
5. The method of claim 1, wherein when screening unstructured text expressions in combination with knowledge of the structured remote sensing image, the screening is based on whether keywords obtained from unstructured text expressions by the target feature can be represented by labeling in the structured remote sensing image.
6. The method of claim 1, wherein the training the remote sensing image semantic environmental model to distributively characterize the target characteristics in the remote sensing image comprises:
randomly acquiring a structured remote sensing image set, and labeling a target represented in the structured remote sensing image set by a keyword obtained by unstructured text expression of the target characteristic to obtain a target characteristic vector;
superposing the target characteristic vector and the word vector corresponding to the keyword, and training the neuro-language model to obtain a remote sensing image semantic environment model;
and carrying out distributed characterization on the target characteristics in the original structured remote sensing image set by using the remote sensing image semantic environment model.
7. The method of claim 6, wherein labeling the targets represented in the set of structured remote sensing images by keywords derived from the target characteristics via unstructured text expressions comprises: the time and space attributes of the target specifically comprise a target name, a target ID, image shooting time, a longitude and latitude position of a target center, four corner points of a minimum circumscribed rectangular frame of the target, a target slice size, spatial resolution and background, and are represented by vectors.
8. A distributed representation system of remote sensing image target characteristics, wherein the distributed representation system of remote sensing image target characteristics utilizes the distributed representation method of remote sensing image target characteristics according to any one of claims 1-7, comprising:
the text characterization module is used for expressing the target characteristics by using prior information of unstructured text;
the screening module is used for screening unstructured text expression by combining the knowledge of the structured remote sensing image;
the distributed characterization module is used for training a semantic environment model of the remote sensing image and characterizing target characteristics in the remote sensing image in a distributed manner.
9. An electronic device comprising a memory and a processor, the memory having stored therein a computer program which, when executed by the processor, causes the processor to implement the method of any of claims 1-7.
10. A computer readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the method according to any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310376114.5A CN116385600B (en) | 2023-04-10 | 2023-04-10 | Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310376114.5A CN116385600B (en) | 2023-04-10 | 2023-04-10 | Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN116385600A true CN116385600A (en) | 2023-07-04 |
CN116385600B CN116385600B (en) | 2023-12-19 |
Family
ID=86963115
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310376114.5A Active CN116385600B (en) | 2023-04-10 | 2023-04-10 | Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116385600B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113298151A (en) * | 2021-05-26 | 2021-08-24 | 中国电子科技集团公司第五十四研究所 | Remote sensing image semantic description method based on multi-level feature fusion |
CN114564586A (en) * | 2022-03-04 | 2022-05-31 | 中信银行股份有限公司 | Unstructured sensitive data identification method and system |
CN114638914A (en) * | 2022-03-21 | 2022-06-17 | 腾讯科技(深圳)有限公司 | Image generation method and device, computer equipment and storage medium |
CN115455504A (en) * | 2022-08-25 | 2022-12-09 | 万翼科技有限公司 | Method, device and equipment for identifying size label and storage medium |
CN115712740A (en) * | 2023-01-10 | 2023-02-24 | 苏州大学 | Method and system for multi-modal implication enhanced image text retrieval |
-
2023
- 2023-04-10 CN CN202310376114.5A patent/CN116385600B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113298151A (en) * | 2021-05-26 | 2021-08-24 | 中国电子科技集团公司第五十四研究所 | Remote sensing image semantic description method based on multi-level feature fusion |
CN114564586A (en) * | 2022-03-04 | 2022-05-31 | 中信银行股份有限公司 | Unstructured sensitive data identification method and system |
CN114638914A (en) * | 2022-03-21 | 2022-06-17 | 腾讯科技(深圳)有限公司 | Image generation method and device, computer equipment and storage medium |
CN115455504A (en) * | 2022-08-25 | 2022-12-09 | 万翼科技有限公司 | Method, device and equipment for identifying size label and storage medium |
CN115712740A (en) * | 2023-01-10 | 2023-02-24 | 苏州大学 | Method and system for multi-modal implication enhanced image text retrieval |
Non-Patent Citations (1)
Title |
---|
刘扬: "《基于MNCC模型的高分辨率遥感影像目标识别》" * |
Also Published As
Publication number | Publication date |
---|---|
CN116385600B (en) | 2023-12-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106649818B (en) | Application search intention identification method and device, application search method and server | |
US9870516B2 (en) | Hand-drawn sketch recognition | |
US10043231B2 (en) | Methods and systems for detecting and recognizing text from images | |
Wilkinson et al. | Neural Ctrl-F: segmentation-free query-by-string word spotting in handwritten manuscript collections | |
Ju et al. | An efficient method for document categorization based on word2vec and latent semantic analysis | |
CN113254655B (en) | Text classification method, electronic device and computer storage medium | |
CN109446299B (en) | Method and system for searching e-mail content based on event recognition | |
CN112559747A (en) | Event classification processing method and device, electronic equipment and storage medium | |
Tallapragada et al. | Improved Resume Parsing based on Contextual Meaning Extraction using BERT | |
CN112307048B (en) | Semantic matching model training method, matching method, device, equipment and storage medium | |
CN115248890A (en) | User interest portrait generation method and device, electronic equipment and storage medium | |
CN113569018A (en) | Question and answer pair mining method and device | |
CN112434533A (en) | Entity disambiguation method, apparatus, electronic device, and computer-readable storage medium | |
CN113392305A (en) | Keyword extraction method and device, electronic equipment and computer storage medium | |
US20170337484A1 (en) | Scalable web data extraction | |
CN116385600B (en) | Distributed characterization method and system for target characteristics of remote sensing image and electronic equipment | |
CN113656556B (en) | Text feature extraction method and knowledge graph construction method | |
CN113468890B (en) | Sedimentology literature mining method based on NLP information extraction and part-of-speech rules | |
CN114579876A (en) | False information detection method, device, equipment and medium | |
CN114003706A (en) | Keyword combination generation model training method and device | |
CN112149389A (en) | Resume information structured processing method and device, computer equipment and storage medium | |
Liu et al. | A novel text classification method for emergency event detection on social media | |
CN114896436B (en) | Network structure searching method based on characterization mutual information | |
CN109325126B (en) | Method and device for objectification processing of language text and computer storage medium | |
CN117216250A (en) | Concept word screening method, device, computer equipment and storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |