Mail processing method based on intelligent glasses

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a mail processing method based on intelligent glasses, and the method comprises the steps: judging whether to determine mailbox account authorization information and obtain an access right based on a wearing state, carrying out the feature extraction based on a mail service target image and an image recognition model, and carrying out the feature extraction; determining an isolation strategy and a sorting strategy based on the unread mail data set, determining a vibration level of the smart glasses according to the isolation strategy, determining broadcast information according to the sorting strategy and obtaining selected mails, performing data processing on the selected mails to determine target selected mails, extracting mail keywords of the target selected mails, and sending the mail keywords to the smart glasses; and determining a playing level based on the mail keyword, judging whether to encrypt the processed mail based on the playing level, and determining whether mail processing is completed according to the relationship between the encrypted hash value and the cloud hash value. According to the invention, the mail processing efficiency and reliability are guaranteed.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and more specifically, to an email processing method based on smart glasses. Background Technology

[0002] With the continuous development of information technology, email, as an important means of information exchange, has been widely used in daily work and life. However, for visually impaired individuals, traditional email reading and processing still presents certain inconveniences. Currently, visually impaired individuals mainly rely on the interaction between Braille readers and email servers to receive and process emails. This linear processing mode is inefficient when processing emails with complex structures or large amounts of information, especially in filtering emails from blacklists and processing key information in lengthy emails, which usually consumes a lot of time, thus affecting email processing efficiency. Furthermore, existing technologies lack the ability to semantically understand and structurally present email content, making it difficult for visually impaired individuals to complete email processing within fragmented time, thereby limiting their information acquisition efficiency and independent processing capabilities. Therefore, it is necessary to design an email processing method based on smart glasses to solve the problems existing in the current technology. Summary of the Invention

[0003] In view of this, the present invention proposes an email processing method based on smart glasses, which aims to solve the problems that usually require a lot of time to filter emails from blacklists and process key information in long emails, thus affecting the efficiency of email processing, as well as the lack of semantic understanding and structured presentation capabilities of email content, making it impossible to effectively extract key information from emails, and making it difficult for visually impaired people to complete email processing in fragmented time.

[0004] This invention proposes an email processing method based on smart glasses, comprising: The wearing status of the smart glasses is determined based on skin contact sensing. Based on the wearing status, it is determined whether to determine the email account authorization information and obtain access permission. Based on the access permission, the initial image of the email service is obtained. The initial image of the email service is preprocessed to determine the target image of the email service. Feature extraction is performed based on the target image of the email service and the image recognition model. An unread email dataset is determined based on the feature extraction results. An isolation strategy and a sorting strategy are determined based on the unread email dataset. The vibration level of the smart glasses is determined based on the isolation strategy. The broadcast information is determined based on the sorting strategy and the selected email is obtained. The selected emails are processed to determine the target selected emails. The email keywords of the target selected emails are extracted, and the playback level is determined based on the email keywords. After the playback level is determined, the target selected emails are marked as processed emails. Based on the playback level, determine whether to encrypt the processed email, determine the encrypted hash value of the encrypted processed email, upload the encrypted processed email or the unencrypted processed email to the cloud database of the smart glasses, determine the cloud hash value of the encrypted processed email, and determine whether the email processing is completed based on the relationship between the encrypted hash value and the cloud hash value.

[0005] Furthermore, when determining whether to confirm email account authorization information and obtain access permissions based on the wearing status, the process includes: The wearing status includes standard wearing status and non-standard wearing status; When the wearing status is the standard wearing status, the email account authorization information is determined and access is obtained; when the wearing status is the non-standard wearing status, the email account authorization information is not determined. When determining the email account authorization information and obtaining the access permission, the smart glasses obtain an access token through the OAuth2.0 protocol and initiate an access request to the mail server based on the access token.

[0006] Furthermore, when preprocessing the initial image for the email service to determine the target image for the email service, the process includes: The initial image for the email service is denoised and converted to a new format. The format conversion converts the denoised initial image for the email service into a PNG or JPEG format in the RGB color space. The result of the format conversion is then normalized in size using bilinear interpolation or adjacent interpolation. The result of the size normalization is then processed in image processing, which includes adjusting contrast and color balance. The target image for the email service is determined based on the result of the image processing.

[0007] Furthermore, when performing feature extraction based on the target image of the email service and the image recognition model, and determining the unread email dataset based on the feature extraction results, the process includes: The target image of the email service is subjected to feature extraction based on a pre-trained convolutional neural network to determine the image feature vector, and the email text data corresponding to the image feature vector is extracted based on a language representation model to construct the unread email dataset. Email text data includes unread emails, unread senders, unread titles, and unread timestamps.

[0008] Furthermore, when determining the isolation and sorting strategies based on the unread email dataset, the following steps are included: When there is an unread sender in the unread email dataset that is the same as the blacklist, the unread emails corresponding to that unread sender in the unread email dataset are determined as the isolation strategy. When there is no unread sender in the unread email dataset that is the same as the blacklist, all unread emails in the unread email dataset are determined as the sorting strategy.

[0009] Furthermore, when determining the vibration level of the smart glasses according to the isolation strategy, determining the broadcast information according to the sorting strategy, and retrieving the selected email, the process includes: The isolation strategy includes: obtaining the number of unread emails from unread senders in the unread email dataset that match the blacklist set, and deleting the corresponding unread emails; When the number of unread messages is greater than or equal to the unread message threshold, the vibration level of the smart glasses is determined to be Level 1 vibration. When the number of unread messages is less than the unread message threshold, the vibration level of the smart glasses is determined to be Level 2 vibration. The vibration amplitude of Level 1 vibration is greater than that of Level 2 vibration. The sorting strategy includes: obtaining all unread emails and their corresponding unread senders, and sorting them in descending order of frequency of occurrence. If there are emails with the same frequency, they are sorted according to the first letter of the unread sender's name. The sorting results are then read aloud, and the selected email is determined according to the selection command.

[0010] Furthermore, when processing the selected emails to determine the target selected email, the process includes: The selected emails are processed based on NLP analysis. The data processing includes word segmentation, stop word removal and stemming. Dependency parsing is used to analyze the relationships between words in the selected emails after data processing, extract email content phrases, and use the email content phrases that conform to the syntactic structure as the email keywords. The selected email after data processing is identified as the target selected email.

[0011] Furthermore, when extracting email keywords from the selected target emails and determining the playback level based on the email keywords, the process includes: The total number of keywords in all emails is counted, and the number of keywords for each email keyword is determined. The ratio of the number of keywords for each email keyword to the total number of keywords is determined. When the ratio is greater than or equal to a threshold, the playback level is determined to be Level 1 broadcast. When the ratio is less than the threshold, the playback level is determined to be Level 2 broadcast. The sound intensity of Level 1 broadcast is greater than that of Level 2 broadcast.

[0012] Furthermore, when determining whether to encrypt the processed email based on the playback level, the process includes: When the playback level is the first-level broadcast, it is determined that the processed email will be encrypted; when the playback level is the second-level broadcast, it is determined that the processed email will not be encrypted.

[0013] Furthermore, when determining whether email processing is complete based on the relationship between the encrypted hash value and the cloud hash value, the process includes: When the encrypted hash value is the same as the cloud hash value, the email processing is considered complete. When the encrypted hash value is different from the cloud hash value, the cloud database uploaded to the smart glasses is considered abnormal, and the encrypted processed email is re-uploaded to the cloud database of the smart glasses.

[0014] Compared with existing technologies, the advantages of this invention are as follows: It automatically determines the wearing status of smart glasses through skin contact sensing, achieving barrier-free activation and identity recognition. By combining image recognition models to extract features from the target service image of emails, it ensures the stability and reliability of emails for visually impaired individuals without requiring them to actively read them, improving their information acquisition efficiency during fragmented time periods. Based on a constructed unread email dataset and the formulation of isolation and sorting strategies, it effectively filters emails from the blacklist, prioritizing the playback of important emails and preventing visually impaired individuals from being overwhelmed by redundant emails. Simultaneously, it intelligently determines the playback level based on email keywords and provides corresponding vibration level feedback and voice playback through the smart glasses, enhancing the intuitiveness and interactivity of email processing. The cloud-based hash value verification mechanism improves the integrity and accuracy of email processing, ensuring the stability and reliability of email processing for visually impaired individuals. Attached Figure Description

[0015] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1A flowchart illustrating an email processing method based on smart glasses, provided as an embodiment of the present invention. Detailed Implementation

[0016] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0017] In some embodiments of this application, see Figure 1 As shown, an email processing method based on smart glasses includes: S100: Determines the wearing status of the smart glasses based on skin contact sensing, determines whether to confirm email account authorization information and obtain access based on the wearing status, obtains the initial image of the email service based on the access permission, and preprocesses the initial image of the email service to determine the target image of the email service.

[0018] S200: Based on the target image of the email service and the image recognition model, feature extraction is performed. The unread email dataset is determined according to the feature extraction results. The isolation strategy and sorting strategy are determined based on the unread email dataset. The vibration level of the smart glasses is determined according to the isolation strategy. The broadcast information is determined according to the sorting strategy and the selected email is obtained.

[0019] S300: Process the selected emails to determine the target emails, extract the email keywords of the target emails, determine the playback level based on the email keywords, and mark the target emails as processed emails after the playback level is determined.

[0020] S400: Determines whether to encrypt processed emails based on playback level, and determines the encrypted hash value of the processed emails. Uploads the encrypted or unencrypted processed emails to the cloud database of the smart glasses, and determines the cloud hash value of the encrypted processed emails. Based on the relationship between the encrypted hash value and the cloud hash value, determines whether to complete email processing.

[0021] Specifically, when visually impaired individuals wear smart glasses, the glasses, through pressure sensors and other detection devices, can sense the contact between the skin and the glasses in real time, dynamically determining the wearing status. Based on this status, it determines whether to authorize email account information and grant access, avoiding resource waste caused by accidental touches or misoperations when the glasses are not worn. Based on access permissions, the smart glasses are used to obtain an initial image of the email service. This initial image reflects all emails for the visually impaired user. Environmental noise or network fluctuations can interfere with the initial image acquisition; preprocessing this image ensures its stability and reliability, laying the foundation for subsequent processing. Feature extraction is performed based on the email service target image and an image recognition model. The image recognition model performs multi-scale feature extraction on the email service target image and combines it with semantic analysis and other techniques to determine the unread email dataset. This dataset includes unread contacts and unread emails, ensuring the comprehensiveness and accuracy of the unread email data for the visually impaired user. The system dynamically determines isolation and sorting strategies based on the number of unread contacts in the unread email dataset. The isolation strategy maps different vibration levels, allowing visually impaired individuals to perceive the isolation status of unread emails in real time based on vibration. The sorting strategy determines the broadcast information, which is then broadcast through the smart glasses' speakers and other devices. This ensures the priority of unread email processing for visually impaired individuals, enabling them to complete email processing within fragmented time. When visually impaired individuals select unread emails, voice and other methods assist them in selecting the emails, ensuring the efficiency of information acquisition and their autonomous processing capabilities.

[0022] Understandably, after the visually impaired group selects an email, the content of that email undergoes data processing, transforming it into a multi-layered structured process. This allows them to grasp the email's keywords and determine the playback level based on those keywords. By integrating voice and visual processing, the time spent filtering and reading emails is reduced, improving efficiency in handling fragmented time. The playback level determination process ensures consistency between the playback level and encryption. AES-256 technology is used to encrypt processed emails, generating an encrypted hash value, which is simultaneously uploaded to a cloud database. The cloud hash value is also recorded, and comparing the encrypted hash value with the cloud hash value ensures the stability of unread email processing. Through a combination of preprocessing, feature extraction, isolation or sorting strategies, and verification, the system not only meets the visually impaired group's dual needs for efficiency and self-management in email processing but also ensures the reliability of privacy protection for processed emails.

[0023] In some embodiments of this application, when determining whether to confirm email account authorization information and obtain access based on the wearing status, the following steps are taken: the wearing status includes a standard wearing status and a non-standard wearing status. When the wearing status is a standard wearing status, it is determined that email account authorization information is confirmed and access is obtained. When the wearing status is a non-standard wearing status, it is determined that email account authorization information is not confirmed. When email account authorization information is confirmed and access is obtained, the smart glasses obtain an access token through the OAuth2.0 protocol and initiate an access request to the mail server based on the access token.

[0024] Specifically, by distinguishing between standard and non-standard wearing states, the standard wearing state indicates that the smart glasses are worn correctly. Only when the smart glasses are worn correctly and make accurate contact with the user's skin will the acquisition of email account authorization and access token be triggered, thereby initiating network communication and access token interaction. This lays the foundation for obtaining the initial image of the email service and avoids the waste of resources caused by accidental touches or misoperations when the glasses are not worn.

[0025] In some embodiments of this application, when preprocessing the initial image of the email service to determine the target image of the email service, the process includes: denoising and format conversion of the initial image of the email service; the format conversion converts the denoised initial image of the email service into PNG or JPEG format in RGB color space; the result of the format conversion is subjected to size normalization processing, which is bilinear interpolation or adjacent interpolation; the result of the size normalization processing is subjected to image processing, which includes adjusting contrast and color balance; and the target image of the email service is determined based on the result of the image processing.

[0026] Specifically, by denoising the initial image of the email service, noise and compression artifacts are effectively removed, ensuring the clarity of the email interface. The format of the denoised initial image of the email service is uniformly converted to PNG or JPEG format in RGB color space, making the converted format compatible with most image input sources, thus simplifying image decoding. Bilinear interpolation or nearest neighbor interpolation is used to normalize the size of the format-converted image. After unifying the resolution, batch parallel processing can be performed, which helps to improve the image quality and visual effect, and reduces image interference caused by resolution differences. The size normalization result is then subjected to contrast enhancement and color balance correction, which helps to highlight text edges and key interface elements, laying the foundation for subsequent feature extraction. This not only improves the accuracy of subsequent recognition, but also takes into account the image quality of the target image of the email service, ensuring the stability of email processing.

[0027] In some embodiments of this application, when performing feature extraction based on the target image of the email service and the image recognition model, and determining the unread email dataset based on the feature extraction results, the process includes: performing feature extraction on the target image of the email service based on a pre-trained convolutional neural network to determine the image feature vector, and extracting the email text data corresponding to the image feature vector based on a language representation model to construct the unread email dataset. The email text data includes unread emails, unread senders, unread titles, and unread timestamps.

[0028] Specifically, by using a pre-trained convolutional neural network to extract deep features from the target image of the email service, visual information such as unread markers, icons, and text positions can be accurately captured, thus ensuring the stability of the image feature vectors. Combined with the BERT (Browser Representation Technology) model, the corresponding email text data (including unread emails, sender information, subject, and timestamp) is parsed, ensuring the completeness and accuracy of the unread email dataset. The process of determining image feature vectors based on convolutional neural networks and determining email text data based on the BERT model (similar to the process of determining text) is mature and lengthy, and will not be elaborated upon here. Convolutional neural networks can extract visual features from the target image of the email service, while the BERT model can perform text extraction in a lightweight deployment environment. The parallel execution of both fully utilizes computing resources and network bandwidth, laying the foundation for subsequent isolation and sorting strategies, and ensuring the stability and reliability of email processing.

[0029] In some embodiments of this application, when determining the isolation strategy and the sorting strategy based on the unread email dataset, the following steps are taken: when there is an unread sender in the unread email dataset that is the same as the blacklist, the unread emails corresponding to that unread sender in the unread email dataset are determined as the isolation strategy; when there is no unread sender in the unread email dataset that is the same as the blacklist, all unread emails in the unread email dataset are determined as the sorting strategy.

[0030] Specifically, by comparing the unread email dataset with a predefined blacklist, the system identifies and classifies unread senders from the blacklist within the unread email dataset, achieving an organic combination of isolation and ranking strategies. When an unread sender matching the blacklist exists in the unread email dataset, the system automatically identifies the corresponding unread email as an isolation strategy, preventing interference or security risks to the mailbox environment of visually impaired users. When the unread email dataset does not contain unread senders from the blacklist, all unread emails are ranked according to the ranking strategy. This not only effectively filters spam, phishing, and other malicious emails, improving mailbox security, but also improves the efficiency of email processing for visually impaired users by determining the ranking strategy based on all unread emails, ensuring the stability and reliability of email processing.

[0031] In some embodiments of this application, when determining the vibration level of the smart glasses according to the isolation strategy, determining the broadcast information according to the sorting strategy, and obtaining the selected email, the following steps are taken: The isolation strategy includes: obtaining the number of unread emails corresponding to unread senders in the unread email dataset that are consistent with the blacklist set, and deleting the corresponding unread emails; when the number of unread emails is greater than or equal to the unread number threshold, the vibration level of the smart glasses is determined to be Level 1 vibration; when the number of unread emails is less than the unread number threshold, the vibration level of the smart glasses is determined to be Level 2 vibration, and the vibration amplitude of Level 1 vibration is greater than that of Level 2 vibration; the sorting strategy includes: obtaining the unread senders corresponding to all unread emails, and arranging them in descending order of the frequency of occurrence of the unread senders; if there are unread senders with the same frequency, arranging them according to the first letter of the unread senders; broadcasting the sorting results by voice; and determining the selected email according to the selection instruction.

[0032] Specifically, the isolation strategy can count the number of unread emails from all unread senders in the unread email dataset that are from the blacklist, and delete these unread emails to ensure the stability of the isolation process. This not only effectively blocks spam, phishing emails, and other potential security threats, avoiding interference with visually impaired individuals, but also reduces the burden of email processing for them. Based on a set unread count threshold, the smart glasses can dynamically adjust the vibration level. When the number of unread emails is greater than or equal to the threshold, the vibration level of the smart glasses is set to Level 1 (large amplitude) to alert visually impaired individuals. When the number of unread emails is less than the threshold, the vibration level of the smart glasses is set to Level 2 (small amplitude), ensuring the flexibility and adaptability of email processing. The sorting strategy retrieves all unread emails from senders without isolation, or from all unread email senders remaining in the non-blacklist after isolation. These senders are then sorted in descending order of frequency, and further refined by alphabetical order of frequency in cases of equal frequency. This ensures the accuracy and stability of the sorting results. By broadcasting the sorting results via voice, visually impaired individuals can select emails to process simply by issuing voice commands, improving the efficiency of email processing during fragmented time and ensuring the stability and reliability of email processing.

[0033] In some embodiments of this application, when determining the target selected email by data processing of selected emails, the process includes: performing data processing on the selected emails based on NLP analysis, including word segmentation, removal of stop words and stemming, using dependency parsing to parse the relationships between words in the selected emails after data processing, extracting email content phrases, using email content phrases that conform to the syntactic structure as email keywords, and determining the selected emails after data processing as the target selected emails.

[0034] Specifically, data processing of selected emails involves three main steps: 1. Word segmentation: Continuous text in the selected email is divided into individual words or phrases to ensure the stability of subsequent processing. 2. Stop word removal: Commonly used words that do not affect the actual semantics of the selected email are removed to improve the effectiveness and reliability of the analysis. 3. Stemming: Words in the selected email are restored to their root forms, such as unifying "performance" and "performance's" into "performance" to reduce semantic redundancy. Dependency parsing is used to analyze the grammatical and semantic relationships between words, extracting noun phrases from the content of the selected email. Dependency parsing can identify noun phrases such as "project development" and "material quotation," which reflect the theme of the selected email and are key to email keyword extraction. By combining NLP technology and dependency parsing, the stability of email keyword extraction and identification is ensured. The extraction of email keywords is based on grammatical structure, ensuring the semantic value of the extracted content and guaranteeing the stability and reliability of email processing.

[0035] In some embodiments of this application, when extracting email keywords from the target selected email and determining the playback level based on the email keywords, the process includes: counting the total number of keywords for all email keywords, determining the number of keywords for each email keyword, determining the ratio of the number of each keyword to the total number of keywords, and determining the playback level as Level 1 broadcast when the ratio is greater than or equal to a threshold value, and determining the playback level as Level 2 broadcast when the ratio is less than the threshold value. The sound intensity of Level 1 broadcast is greater than that of Level 2 broadcast.

[0036] Specifically, by statistically analyzing the number and ratio of keywords in selected emails, intelligent classification of playback levels is achieved. The total number of keywords in all emails is counted, and the number of keywords for each email is determined. The ratio of each keyword's number to the total number of keywords is also calculated. For example, if "project development" and "material quotation" are email keywords, and "project development" appears twenty times in the selected emails while "material quotation" appears ten times, the total number of keywords is two, the ratio for "project development" is ten, and the ratio for "material quotation" is five. A comparison is made between the maximum ratio and a threshold ratio. When the ratio is greater than or equal to the threshold, it indicates that the selected emails cover a comprehensive range of keywords and are of high importance. In this case, the playback level is determined to be Level 1, with a higher volume to draw the attention of visually impaired users to the important selected emails. Conversely, when the ratio is less than the threshold, the playback level is determined to be Level 2, with a lower volume to draw the attention of visually impaired users. This ensures the flexibility and adaptability of email processing.

[0037] In some embodiments of this application, when determining whether to encrypt processed emails based on playback level, the method includes: when the playback level is level 1, it is determined that processed emails should be encrypted; when the playback level is level 2, it is determined that processed emails should not be encrypted.

[0038] Specifically, when the playback level is Level 1, the processed email content is encrypted to effectively prevent important processed emails from being stolen or tampered with by unauthorized parties during transmission and uploading. When the playback level is Level 2, it indicates that the processed emails are of lower importance, so encryption is waived and they can be directly uploaded to the cloud database of the smart glasses. By flexibly determining the relationship between playback level and encryption mapping, a balance between security strategy and resources is achieved, ensuring the intelligence and reliability of email processing.

[0039] In some embodiments of this application, when determining whether email processing is complete based on the relationship between the encrypted hash value and the cloud hash value, the following steps are taken: when the encrypted hash value and the cloud hash value are the same, it is determined that the email processing is complete; when the encrypted hash value and the cloud hash value are different, it is determined that there is an anomaly in the cloud database uploaded to the smart glasses, and the encrypted processed email is re-uploaded to the cloud database of the smart glasses.

[0040] Specifically, before uploading to the smart glasses' cloud database, processed emails that require encryption are encrypted using AES-256 technology to generate a cryptographic hash value. This hash value is a unique digital representation of the processed email's content, ensuring that the email has not been tampered with during transmission and storage. After uploading to the smart glasses' cloud database, the cloud hash value of the stored processed email is determined again. This cloud hash value is a digital representation of the actual content of the processed email stored in the cloud database. The cryptographic hash value is compared with the cloud hash value. If these two values ​​match, it means the processed email has not changed during transmission and storage, indicating that email processing is complete. If the cryptographic hash value and the cloud hash value do not match, it indicates that the processed email may have been lost, damaged, or tampered with during the upload process. At this point, an anomaly is detected in the cloud database uploaded to the smart glasses. The encrypted processed email needs to be re-uploaded to the cloud database of the smart glasses. By checking the consistency of the hash value, the reliability of the completed email processing is ensured, guaranteeing the efficiency of information acquisition and self-processing ability of the visually impaired community. Furthermore, when it is necessary to trace back and retrieve the email, the reliability and stability of the email processing are guaranteed.

[0041] In summary, the beneficial effects of this invention are as follows: It automatically determines the wearing status of smart glasses through skin contact sensing, achieving barrier-free activation and identity recognition. By combining image recognition models to extract features from the target service image of emails, it ensures the stability and reliability of email access for visually impaired individuals without requiring active reading, improving their information acquisition efficiency during fragmented time periods. Based on the constructed unread email dataset and the formulation of isolation and sorting strategies, it effectively filters emails from the blacklist, prioritizing the playback of key emails and preventing visually impaired individuals from being overwhelmed by redundant emails. Simultaneously, it intelligently determines the playback level based on email keywords and provides corresponding vibration level feedback and voice playback through the smart glasses, enhancing the intuitiveness and interactivity of email processing. The cloud-based hash value verification mechanism improves the integrity and accuracy of email processing, ensuring the stability and reliability of email processing for visually impaired individuals.

[0042] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for processing mail based on smart glasses, characterized in that, include: The wearing status of the smart glasses is determined based on skin contact sensing. Based on the wearing status, it is determined whether to determine the email account authorization information and obtain access permission. Based on the access permission, the initial image of the email service is obtained. The initial image of the email service is preprocessed to determine the target image of the email service. Feature extraction is performed based on the target image of the email service and the image recognition model. An unread email dataset is determined based on the feature extraction results. An isolation strategy and a sorting strategy are determined based on the unread email dataset. The vibration level of the smart glasses is determined based on the isolation strategy. The broadcast information is determined based on the sorting strategy and the selected email is obtained. The selected emails are processed to determine the target selected emails. The email keywords of the target selected emails are extracted, and the playback level is determined based on the email keywords. After the playback level is determined, the target selected emails are marked as processed emails. Based on the playback level, determine whether to encrypt the processed email, determine the encrypted hash value of the encrypted processed email, upload the encrypted processed email or the unencrypted processed email to the cloud database of the smart glasses, determine the cloud hash value of the encrypted processed email, and determine whether the email processing is completed based on the relationship between the encrypted hash value and the cloud hash value.

2. The smart glasses based mail processing method of claim 1, wherein, When determining whether to confirm email account authorization information and obtain access permissions based on the wearing status, the following are included: The wearing status includes standard wearing status and non-standard wearing status; When the wearing status is the standard wearing status, the email account authorization information is determined and access is obtained; when the wearing status is the non-standard wearing status, the email account authorization information is not determined. When determining the email account authorization information and obtaining the access permission, the smart glasses obtain an access token through the OAuth2.0 protocol and initiate an access request to the mail server based on the access token. 3.The smart glasses based mail processing method of claim 2, wherein, When preprocessing the initial image of the email service to determine the target image of the email service, the process includes: The initial image for the email service is denoised and converted to a new format. The format conversion converts the denoised initial image for the email service into a PNG or JPEG format in the RGB color space. The result of the format conversion is then normalized in size using bilinear interpolation or adjacent interpolation. The result of the size normalization is then processed in image processing, which includes adjusting contrast and color balance. The target image for the email service is determined based on the result of the image processing. 4.The smart glasses based mail processing method of claim 3, wherein, When performing feature extraction based on the target image of the email service and the image recognition model, and determining the unread email dataset based on the feature extraction results, the process includes: Feature extraction is performed on the target image of the email service based on a pre-trained convolutional neural network to determine the image feature vector, and the email text data corresponding to the image feature vector is extracted based on a language representation model to construct the unread email dataset. Email text data includes unread emails, unread senders, unread titles, and unread timestamps. 5.The smart glasses based mail processing method of claim 4, wherein, When determining the isolation and sorting strategies based on the unread email dataset, the following are included: When there is an unread sender in the unread email dataset that is the same as the blacklist, the unread emails corresponding to that unread sender in the unread email dataset are determined as the isolation strategy. When there is no unread sender in the unread email dataset that is the same as the blacklist, all unread emails in the unread email dataset are determined as the sorting strategy.

6. The email processing method based on smart glasses according to claim 5, characterized in that, When determining the vibration level of the smart glasses according to the isolation strategy, determining the broadcast information according to the sorting strategy, and retrieving the selected email, the process includes: The isolation strategy includes: obtaining the number of unread emails from unread senders in the unread email dataset that match the blacklist set, and deleting the corresponding unread emails; When the number of unread messages is greater than or equal to the unread message threshold, the vibration level of the smart glasses is determined to be Level 1 vibration. When the number of unread messages is less than the unread message threshold, the vibration level of the smart glasses is determined to be Level 2 vibration. The vibration amplitude of Level 1 vibration is greater than that of Level 2 vibration. The sorting strategy includes: obtaining all unread emails and their corresponding unread senders, and sorting them in descending order of frequency of occurrence. If there are emails with the same frequency, they are sorted according to the first letter of the unread sender's name. The sorting results are then read aloud, and the selected email is determined according to the selection command. 7.The smart glasses based mail processing method of claim 6, wherein, When processing the selected emails to determine the target selected emails, the process includes: The selected emails are processed based on NLP analysis. The data processing includes word segmentation, stop word removal and stemming. Dependency parsing is used to analyze the relationships between words in the selected emails after data processing, extract email content phrases, and use the email content phrases that conform to the syntactic structure as the email keywords. The selected email after data processing is identified as the target selected email. 8.The smart glasses based mail processing method of claim 7, wherein, When extracting email keywords from the selected target emails and determining the playback level based on the email keywords, the process includes: The total number of keywords in all emails is counted, and the number of keywords for each email keyword is determined. The ratio of the number of keywords for each email keyword to the total number of keywords is determined. When the ratio is greater than or equal to a threshold, the playback level is determined to be Level 1 broadcast. When the ratio is less than the threshold, the playback level is determined to be Level 2 broadcast. The sound intensity of Level 1 broadcast is greater than that of Level 2 broadcast. 9.The smart glasses based mail processing method of claim 8, wherein, When determining whether to encrypt the processed email based on the playback level, the following steps are included: When the playback level is the first-level broadcast, it is determined that the processed email will be encrypted; when the playback level is the second-level broadcast, it is determined that the processed email will not be encrypted. 10.The smart glasses based mail processing method of claim 9, wherein, When determining whether to complete email processing based on the relationship between the encrypted hash value and the cloud hash value, the process includes: When the encrypted hash value is the same as the cloud hash value, the email processing is considered complete. When the encrypted hash value is different from the cloud hash value, the cloud database uploaded to the smart glasses is considered abnormal, and the encrypted processed email is re-uploaded to the cloud database of the smart glasses.

Images