33 results about "Base64" patented technology

The invention discloses a Web interface element recognition method, system and device and a medium, belongs to the technical field of artificial intelligence and Web interface development, and aims to solve the technical problem of how to improve the recognition rate of dynamic elements and CSS hidden elements so as to enable Web interface element positioning to be more accurate. Performing screenshot by using a playwright screenshot function, converting the screenshot into a Base64 coded character string by using a function after the screenshot is completed, inputting the character string and a url into an intelligent agent, calling a DeepSeek-VL2 visual model by the intelligent agent, analyzing the screenshot and outputting identified structured element data; dOM structure analysis: inputting a page url address, and generating DOM structure data by using a DOM analysis function; dOM tree analysis and visual element matching: calling an element matching function, performing visual element and DOM structure matching, and obtaining all successfully matched element data; interface element extraction and positioning generation; and element application.

The invention discloses an RAG document splitting optimization method and system in the operator field, and belongs to the technical field of large model optimizing.The method comprises the steps that a document is uploaded, and source files and file information are stored through minio and paradeDB respectively; constructing a document loader, automatically selecting a corresponding loader according to a file type to carry out file processing, converting the file into a uniform Markdown text, analyzing a picture in the document, and converting the picture into a base64 format; constructing a picture processor, extracting a picture base64 character string in the text, processing a bitmap and a vector diagram, and converting the bitmap and the vector diagram into a Markdown picture reference format; constructing a document divider; vector conversion; and recalling the text. According to the method, the problem that picture extraction, conversion and processing are complex in the current RAG is solved, vector retrieval, full-text retrieval and mixed retrieval functions can be achieved through one database, and the operation and maintenance cost is reduced.

The invention provides an e-commerce commodity graph differentiation generation method and device, equipment and a medium, and the method comprises the steps: obtaining an original to-be-shelved commodity graph of e-commerce, and finally obtaining the disassembly information; the method comprises the following steps: adjusting an original commodity image into 980 * 980 pixels, performing Gaussian blur processing at the same time, and then converting image pixel data into a Base64 coded character string; constructing a composite cue word according to the image data, the disassembly information and a format required by generation; inputting the composite cue word into a model to generate a new scene description; inputting the new scene description, the original commodity graph and the synthesis requirement into a model to generate a synthesis cue word, and then inputting the original commodity and the synthesis cue word into an image-text model to generate a differentiated commodity graph; setting the output resolution, the number of iterations and the learning rate of the SuperIR model, and inputting the differentiated commodity graph into the SuperIR model to generate a differentiated commodity optimization graph; the defects of low efficiency, poor scene adaptation, main body distortion and difficulty in repeated avoidance of a traditional method are overcome.

The application discloses a QR code encryption and decryption method and system based on chaos theory and an AES algorithm, relates to the encryption and decryption technical field of QR codes, and comprises an encryption process and a decryption process; the encryption process comprises the following steps: firstly, chaotic sequences generated by improved Logistic mapping and Chebyshev mapping are subjected to XOR to obtain a composite chaotic sequence, the composite chaotic sequence is used as an initial key of AES grouping, and then ciphertext is obtained through AES encryption. Grouping initial keys are dynamically generated, a newly generated key is used for encrypting one group of plaintext data, the keys are completely independent, and therefore, one-time password is realized. Then, the AES algorithm is improved, plaintext information is first encrypted, then secondary encryption is performed on the plaintext information in a Base64 mode, final ciphertext information is subjected to QR code standard coding, and an encrypted QR code is generated. Through the mode of increasing encryption processing, the application effectively reduces information leakage or information tampering problems caused by QR codes.

The invention relates to the technical field of cross-system data processing, and particularly provides an intelligent clipboard content processing method and system, and the method comprises the following steps: S1, content type detection; s2, Base64 decoding is carried out; s3, performing path conversion; and S4, output control. Compared with the prior art, intelligent conversion and standardized output of clipboard content can be realized through dynamic content identification and multi-format output control.

The invention discloses a high-performance asynchronous segmentation Base64 coding method and device based on a VTE system. The method comprises the following steps: dynamically adjusting the segmentation size according to the data type; the method comprises the following steps: uniformly converting an Array Buffer and a character string into a character string coded by UTF-8 (Untranslated Transform Factor-8) by adopting a preprocessor Data method, and processing illegal characters according to a fault-tolerant strategy at the same time; the method comprises the following steps of: completing single-segment coding by adopting an encoder Segment, and putting coding logic into a micro-task queue through await; calculating and displaying the encoding progress; a yieldMain Thread method is adopted, and a main thread is preferentially made to process key operations or key tasks; therefore, large file coding is prevented from blocking a main thread, and real-time response of key operations such as image preview and medical record editing of the VTE system is guaranteed; unicode medical characters are adapted, illegal character fault tolerance of medical data is supported, it is ensured that the coding process is not interrupted, and complete transmission of the data is guaranteed; medical scene progress feedback is provided, the operation experience of medical staff is improved, and repeated operation is avoided; the method is compatible with a native btoa API, and the transformation and migration cost of a VTE system is reduced.

The invention discloses a method and a system for stably displaying a Base64 picture at a Web end. The server performs Base64 coding on pictures not exceeding 1MB, generates content fingerprints, associates versions with metadata, and distributes description data through homologous signatures; and the front end asynchronously loads and displays after coding and optional zooming on the canvas side, and when the loading fails, hierarchical backspacing taking the placeholder as the first is executed. The system can selectively start a fragmented list and memory water level control, combines local cache and flicker-free replacement with fingerprints and versions as keys, and supports operation rapid picture changing and history tracing. According to the scheme, the problems of cross-domain pollution, first screen blank and memory peak are solved, and the loading stability and the operation and maintenance efficiency are improved.

The invention discloses an unstructured data processing method and device. The method comprises the following steps: establishing a secure data channel between a bank external system and a bank internal unstructured data storage system; based on the security data channel, receiving an access request from a bank external system; processing unstructured data according to the type of the access request; when the access request is a query request, generating an unstructured data identifier based on a request parameter; calling a software toolkit of the unstructured data storage system to pull a corresponding file of the query request to local storage, and forming an unstructured data file and a corresponding return virtual path; receiving an access instruction for the unstructured data file; and when the access request is an uploading request, decoding the Base64 coding file, calling the software toolkit, and storing the software toolkit in an unstructured data storage system. The method and the device are used for improving the efficiency and the accuracy of querying, uploading and maintaining the non-structural data.

This invention discloses an asynchronous segmented Base64 decoding method and system for a hospital-wide integrated clinical rehabilitation system. The method includes: S1: Dynamic segmentation; parsing the Base64 encoded string to identify the data type, and dividing the segment length according to the data type and the total data length; S2: Asynchronous task scheduling; dividing the decoding task into multiple levels according to the priority of medical business, and allocating the decoding task to the core thread pool and elastic thread pool according to the level; S3: Parallel decoding; precompiling the mapping relationship between Base64 encoded characters and their corresponding binary values ​​into a hash table, and using SIMD instructions to perform parallel operations on the segmented encoded data; S4: Fault tolerance verification; verifying integrity; S5: Result merging and callback; thereby improving decoding efficiency and ensuring integrity.

The invention discloses a Base64 invoice automatic read-write optimization method based on artificial intelligence. The method comprises the following steps: S1, executing integrity verification on an electronic invoice file uploaded by a user and generating Base64 coded invoice data; s2, transmitting the Base64 coded invoice data to a cloud service platform through an HTTPS protocol, and executing field structured extraction processing; s3, calling an invoice data interface of the tax system, and performing validity and consistency check on the structured invoice field set; s4, inputting the structured invoice field set into the evidence deep learning model to generate an uncertainty analysis result; s5, respectively calculating a data uncertainty score and a model uncertainty score; s6, judging the relationship between the uncertainty decomposition result and a preset threshold value; and S7, if the uncertainty decomposition results do not exceed the corresponding preset threshold value, marking the structured invoice field set and the tax check result as a credible state. According to the invention, an evidence deep learning model is adopted, and automatic Base64 invoice reading and writing are realized.

A suspicious Uniform Resource Locator (URL) of a resource on the public Internet is detected by searching a part of the URL for a candidate substring, which is the longest possible substring that is in accordance with a Base64 encoding scheme. The candidate substring is converted to a candidate binary data in accordance with the Base64 encoding scheme. The candidate binary data is then converted to a candidate American Standard Code for Information Interchange (ASCII) string. The candidate ASCII string is evaluated to determine the information type of the candidate ASCII string. A determination as to whether the URL is suspicious is based at least on the information type of the candidate ASCII string.

The invention discloses a flattening HTTP (Hyper Text Transport Protocol) decoding method and system based on multi-classification learning and a medium. A multi-classification HTTP coding detection model is generated by constructing an HTTP coding data set, extracting a multi-classification feature matrix of the coding data set and training the coding data set through multi-classification learning; extracting an HTTP original load from the HTTP traffic, extracting a multi-classification feature vector of the HTTP original load, loading a multi-classification HTTP coding detection model, and carrying out multi-classification HTTP coding detection; judging an HTTP coding detection result; if the code of the original load in the HTTP flow is detected, flattening HTTP decoding is carried out; recording and outputting an HTTP decoding result; according to the method, the original load is extracted from the HTTP flow, multi-classification machine learning training detection model is carried out aiming at multiple types of codes such as URL, Base64, Unicode, Hex and HTML, flattening decoding operation is carried out according to a model detection result, and finally the clear HTTP load is cleaned out.

The invention discloses a multi-type HTTP (Hyper Text Transport Protocol) coding feature extraction method and system and a medium. The method comprises the following steps: constructing a multi-type HTTP coding data set; sample data of URL codes, Base64 codes, Unicode codes, Hex codes, HTML codes and combined codes and nested codes of the URL codes, the Base64 codes, the Unicode codes, the Hex codes and the HTML codes are collected respectively, and corresponding type labels are marked; the method comprises the following steps: reading a multi-type HTTP coded data set, sequentially reading samples, and carrying out feature extraction; constructing a multi-type coding feature matrix; constructing a coding sample feature vector, constructing and standardizing a feature matrix of the multi-type HTTP coding data set, constructing a coding type vector matrix according to type values in the multi-type HTTP coding data set, and performing transverse splicing operation on the coding type vector matrix and the standardized feature matrix to construct a multi-type coding feature matrix; according to the method, a multi-type coding data set which can be used for coding learning training, verification and testing is constructed, feature extraction is performed through unique features of multi-type coding, and the accuracy and efficiency of feature extraction are improved.

The application discloses an intelligent canvas interaction system and method based on AIAgent. The system comprises a front-end canvas module, a ChatWebSocket channel, a BridgeWebSocket channel, an AIAgent decision module, a tool registration and routing module, and a front-end tool executor. The system assigns an execution position identifier to each tool in the tool set and defines a routing decision function. The canvas state is synchronized using an incremental superposition mechanism, and the layer set, element set, and configuration parameters are updated according to a formula after the first tool call. The system supports the canvas_state tool for AI to obtain the current canvas state in real time to assist in parameter decision-making, and the canvas_preview tool for performing a screenshot on the canvas and returning a base64 encoded image for AI visual verification. The method comprises the steps of instruction receiving, semantic analysis, tool routing, cross-end execution, result summarization, and response generation. The application realizes natural language driven multi-step canvas operation, effectively reducing the user threshold.

The application provides a kind of compatible DBC's automobile communication method and system in the field of automobile communication, and the method comprises the following steps: step S10, based on DBC format definition DBCE format;Step S20, when automobile communication, the sending end is based on the DBCE format and the CAN message to be transmitted generation DBCE file, and the DBCE file is sent to receiving end after being encrypted using Base64 algorithm;Step S30, the receiving end is decrypted after receiving the DBCE file using Base64 algorithm, and based on the DBCE format, DBCE file is parsed in real time, and then automobile communication is completed.The application has the advantages that: the flexibility, expansibility, efficiency and safety of automobile communication are greatly improved.

The present invention provides a system (100) and method (200) for embedding custom metadata in Over-The-Top (OTT) streaming manifests to enhance the user experience. The system (100) includes an encoder and packager module (108) configured to generate Adaptive Bitrate (ABR) segments and corresponding media manifests. A manifest updater (112) embeds custom metadata within the manifests, including playback control information such as content advisories, scene-specific annotations, and interactive elements. The custom metadata is embedded as XML elements for DASH and UTF-8 tags for HLS, encoded in base64-encoded JSON format. The media manifests with embedded metadata are transmitted to a Content Delivery Network (CDN) (104), which delivers them to client devices (106). The client devices (106) may process the custom metadata or continue streaming without interruption if the metadata is not recognized, ensuring backward compatibility. The method (200) facilitates seamless integration of metadata, enhancing playback without affecting legacy systems.

The application discloses a picture transmission method and device, a picture display method and device, and relates to the technical field of computers.The picture transmission method comprises the following steps: obtaining picture data, a picture encryption key, a transmission encryption key and encrypted interference data; encrypting the picture data by using the picture encryption key to obtain a picture encryption result, inserting the picture encryption key into the picture encryption result to obtain an insertion result, and recording insertion position data; performing base64 encoding on the insertion position data and the encrypted interference data to obtain a processing result; splicing the insertion result and the processing result, encrypting the spliced result by using the transmission encryption key to obtain transmission data; performing base64 encoding on the encryption algorithm name information of the encryption key to obtain an encoding result; and sending the encoding result and the transmission data to a receiving end.In the application, the picture data does not directly participate in data transmission, and the security and privacy of the picture data in the picture transmission process are improved.

The invention provides a Java file hybrid encryption method based on dynamic blocking and compression, and relates to the technical field of computer security, and the method comprises the following steps: generating an RSA key pair, and encrypting and storing the key through Base64 coding; the method comprises the following steps: converting a Java file into binary data, dividing the binary data into a plurality of blocks according to RSA encryption length limitation, respectively carrying out Base64 coding and RSA encryption, generating encrypted blocks, integrating the encrypted blocks into an encrypted file, and generating a digital signature by using a private key to ensure the integrity. In the loading process, automatic decryption and signature verification are conducted through the user-defined class loader embedded with decryption logic, all encryption and decryption operations are achieved through a C + + dynamic link library, interaction with a Java program is achieved through a JNI interface, and safety and execution efficiency are improved.

The invention discloses a Keyblock user name password encryption transmission method and device and a medium, and the method comprises the steps: injecting an RSA public key character string into a front-end login page, encrypting a plaintext password inputted by a user before a form is submitted, and generating a Base64 coded ciphertext; replacing the password field value in the original form with the encrypted ciphertext, and submitting the form carrying the password of the ciphertext to a Keyblock server; the method comprises the following steps: at a Keyblock server side, acquiring a ciphertext password through a self-defined Authority, and decrypting by using a pre-stored RSA private key to obtain a corresponding plaintext password; and re-injecting the decrypted plaintext password into the Keyblock standard authentication process to complete the verification of the user name password. The full-link encryption from the client to the server is ensured, and the risk of application layer plaintext leakage which cannot be protected by the HTTPS protocol is solved.

The invention provides an e-commerce multi-grid commodity graph differentiation generation method and device, equipment and a medium, and the method comprises the steps: obtaining an original e-commerce multi-grid commodity graph, extracting each sub-graph, and recognizing the commodity main body type of each sub-graph; finally, disassembling information is obtained; the size of each sub-image is changed into 980 * 980 pixels, noise is removed, and the sub-images are converted into Base64 codes; constructing a corresponding composite prompt word according to the image data, the disassembly information and a format required by generation; inputting each composite cue word into a model to generate a new scene description corresponding to each sub-graph; inputting the original multi-grid commodity graph, the new scene description corresponding to each sub-graph and the set constraint into a model to generate a replacement prompt word; inputting the replacement prompt word and the original multi-grid commodity graph into a model to generate a multi-grid commodity replacement graph; the defects of scene replacement generalization, complex positioning, style splitting and low efficiency of a traditional method are overcome.

The invention provides a key information picture coding method based on a traditional ocr and a large model. The key information picture coding method comprises the following steps: an environment construction step: deploying image processing, optical character recognition, two-dimensional code / bar code analysis, a large model related tool library and a framework; an image preprocessing step: receiving Base64 coded data of a picture, decoding and converting the Base64 coded data into a standard image format, and further converting the Base64 coded data into an array format; a character recognition and positioning step: carrying out character content recognition and spatial position positioning on the preprocessed image by adopting an optical character recognition tool; a sensitive information detection step: on the basis of the sensitive information identification keyword, carrying out sensitive information detection of multiple types of targets on a character identification result and an image to obtain desensitization information; a sensitive information integration and coding step: integrating detection results of multiple types of sensitive information, and carrying out duplicate removal processing on repeated sensitive information of which the overlapping rate exceeds a preset threshold value in the same region; outputting a result; the problem that the traditional OCR cannot process special pictures is solved.

The invention discloses an MMS multichannel information model transmission scheduling system based on a file IO stream, and relates to the technical field of system architecture design. According to the method, the file content in the MMS model is transmitted in an IO stream format, and the file content information is converted into the Base64 code to be encrypted, so that the limitation of a system UI visual submission model is solved; after the content submitted by the MMS model is obtained and analyzed, a server connection object is created through an SFTP protocol, and connection login is remotely carried out on the file server through an account, an IP, a port and a password; after the file server is logged in based on the SFTP protocol, the file information is converted into the File object and is stored in the formulated path under the file server in a cross-server manner, so that the problem that the memory of the database is occupied when the file is stored in the database is relieved, the controllability of file management is improved, only one model corresponds to multi-channel reporting through a multi-channel concurrent reporting model, and the reliability of file management is improved. Therefore, the model reporting efficiency is improved, and the success rate of issuing the short message by the user is improved.

The invention provides an HTML (Hypertext Markup Language) page local hot update method based on DOM positioning and feature comparison. After a front-end page obtains an HTML source code, the HTML source code is preprocessed to remove blanks, annotations and continuous spaces, compressed through a GZIP algorithm and converted into a character string through BASE64 coding, and the character string is sent to a rear end. A back-end application analyzes a source code to generate a Document object, and extracts front-end head and body feature information, including an element hash value and a node path. During updating, the back end queries the hot update code, analyzes and generates update feature information, and compares the difference between the front and back features to generate an operation instruction set. After the front end receives the instruction, elements are accurately positioned according to node paths, and corresponding operation is executed to complete local updating. The method solves the problems of inaccurate DOM element positioning, overlarge updating range and low transmission efficiency in the existing hot updating technology, realizes on-demand local updating, reduces the network bandwidth occupation, improves the updating efficiency and optimizes the user experience.

The invention relates to a mail content feature matching method and system based on a flow analysis system, which realize non-complete cache stream processing of a subject and a main body of a mail through TCP stream recombination, line-by-line MIME analysis and dynamic code identification, effectively solve the problem of delay of long mail body extraction in the prior art, and improve the extraction efficiency. And the accuracy of continuous matching of the inter-row keywords is ensured. Meanwhile, limited byte cache and Base64 grouping characteristics are utilized to reduce memory occupation, rule grouping transcoding, hit mark priority adjustment and alarm frequency suppression are combined, matching timeliness and system load efficiency are remarkably improved, a prevention and control strategy is taken into effect in time on the premise of ensuring characteristic correctness, missed transmission of malicious mails is avoided, and the service life of the system is prolonged. And an efficient and real-time network security guarantee is provided for a traffic analysis system.

The invention relates to an encryption method of random IV and HEX codes based on AES. The method comprises the following steps: acquiring original plaintext data; a Web Crypto API is adopted to generate a random IV, and through the random IV and a predetermined AES key, an AES encryption algorithm is adopted to encrypt the original plaintext data to obtain a ciphertext; and encoding the ciphertext according to a dynamic HEX-Base64 hybrid encoding method, so as to obtain final encoded data. According to the method, the random IV and AES encryption are combined with the dynamic HEX-Base64 hybrid coding, so that WAF interception is effectively avoided, the false alarm rate is reduced, and the data transmission security and the service continuity are guaranteed.

The invention belongs to the technical field of security and protection equipment encryption, and particularly relates to an encryption method for a debugging port of security and protection equipment, which comprises the following steps: S1, PC (Personal Computer) end tool software generates an encryption key according to information of a unique identification code, a machine type number and a version number; and S2, taking the obtained key Key and the initialization vector IV as a key and an initialization vector of an AES-CBC algorithm respectively, performing encryption operation on an encrypted plaintext to generate a ciphertext in a binary format, then encoding the ciphertext through Base64, converting the ciphertext into a transmittable ciphertext in a character string form, and sending the ciphertext to security and protection equipment. According to the method, the encryption key is dynamically generated by introducing the unique identification information of the equipment and combining the algorithm, so that the security level of the debugging port is remarkably improved, and since the encryption mechanism is based on the unique identification of multiple equipment, the key of other equipment cannot be derived through single equipment, and batch cracking and reverse attack are effectively prevented.

The application provides a method and system for LLM to accurately process long class names, the method comprising: starting a browser, loading a target webpage A of a list to be identified and waiting for the webpage to be completely rendered; injecting a customized JavaScript script into the target webpage A based on a browser debugging protocol, performing layer-by-layer noise reduction processing on an original DOM, and outputting a lightweight DOM string and an original class name and compressed class name mapping table; calling a page screenshot interface through a CDP protocol to perform screenshot, and converting binary data of the screenshot into a Base64 encoded string; constructing an LLM input prompt word containing the lightweight DOM string, the Base64 encoded screenshot and a set task instruction, inputting the LLM to obtain an XPath selector containing the compressed class name; and restoring the XPath selector into an available XPath selector containing the original class name; the method can avoid character expansion caused by random long class names generated by a website, and enable the LLM to obtain relatively complete webpage structure information.

The application discloses a method for generating posters in batches based on Photoshop files, and comprises the following steps: S1, converting Photoshop source file design drawings into browser html display elements in batches; S2, modifying the html display elements of each design drawing in batches; S3, converting each html display element into a base64 picture; and S4, downloading the base64 pictures in batches. The method has the advantages that: through analysis and assembly of source file data, grouping, layering, size and other related data of the source file, the related data can be accurately positioned to the size and position of each element, and the design drawing can be completely restored. Through color integration processing of the source file data, the color can be normally rendered and displayed, the background color of the design drawing can be normally restored, and the user can obtain the style and design of the original design drawing without distortion. Through data driving and algorithm combination, the design drawing elements can be modified in batches, and after modification, all the modified pictures can be downloaded at one time, thereby improving the modification efficiency.

The application discloses a KNN retrieval method of remote sensing images in a visual language model, which comprises five steps: S1, data preprocessing: the remote sensing image data set in the visual language test data set is subjected to base64 processing, and the text data set is subjected to jsonl processing, and the processed data is uploaded to a database; S2, image text serialization: the preprocessed data set is serialized to facilitate random reading during training; S3, CLIP model training: the serialized data is sent to a visual language model for training, and a training model is obtained; S4, image-text feature extraction: the test data is subjected to feature extraction through the training model; and S5, KNN retrieval evaluation: the extracted files are subjected to KNN retrieval evaluation. The application can train the KNN retrieval through the visual language test data set and perform retrieval evaluation, so that relevant or similar content can be quickly and accurately retrieved from massive data.

The invention relates to a method and a device for agile configuration of installation and maintenance based on a digital retina terminal, and belongs to the field of embedding, and the method comprises the following steps: receiving multicast information of an installation and maintenance client through a monitoring multicast mode, and explaining multicast data according to a private protocol; tCP long connection with the installation and maintenance client is established, and meanwhile, heartbeat packet keep-alive with the installation and maintenance client is kept; monitoring and receiving data of the unitized domain name parameters, and decrypting BODY data of the data through key information; performing base64 decoding on the encrypted base64 data of the unitized domain name to obtain actual encrypted unitized domain name data, decrypting the actual encrypted unitized domain name data to obtain decrypted unitized domain name data, and returning the decrypted unitized domain name data to the platform; and generating a receiving entrance domain name parameter mark file, and playing a voice file representing that the agile configuration is successfully installed and maintained. According to the method and the device, the efficiency of scheduling the unitized domain names of the equipment can be effectively improved, the installation and maintenance efficiency and the development efficiency are improved, and the cost and the working difficulty are reduced.