Auxiliary vocabulary memorizing method and system based on deep learning English
Through a deep learning-based method, personalized English vocabulary memory scenarios are generated using user journey and interpersonal relationship data, which solves the problems of low vocabulary memory efficiency and difficulty in stimulating interest in traditional methods, and achieves efficient vocabulary learning results.
Patent Information
- Application Number
- CN202510801282.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-16
- Publication Date
- 2025-10-03
AI Technical Summary
Traditional English vocabulary memorization methods lack connection with real context, have low memory efficiency, cannot be personalized to adapt to users' life trajectories and social relationships, and are difficult to stimulate learning interest.
By obtaining the user's travel data and interpersonal relationship data, an English vocabulary memory scenario that is highly consistent with the real-life scenario is generated. The optimal memory scenario is automatically recommended based on the vocabulary matching number, proficiency and interpersonal relationship data, and a derivative memory scenario is generated through the scenario guidance interface.
It improves the memory efficiency and stickiness of vocabulary learning, strengthens the connection with daily life, solves weak links in a targeted manner, and avoids fatigue from repeated learning.
Smart Images
Figure CN120744221A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of foreign language vocabulary learning, and in particular to a method and system for assisting vocabulary memorization based on deep learning of English. Background Art
[0002] With the continuous acceleration of the process of globalization, English has become a common bridge in international communication, academic research, business cooperation and other fields. Traditional English vocabulary memorization methods mainly rely on classroom teaching, textbook exercises and memory cards, and current English vocabulary memorization methods mainly rely on mechanical repetition or general scene memory APPs. Traditional methods have significant defects: out of real context: vocabulary is presented in an isolated form and lacks connection with actual life scenes, resulting in low memory efficiency and high forgetting rate; lack of personalization: unified learning content cannot adapt to the user's life trajectory and social relationships, and it is difficult to stimulate learning interest; single situation: existing methods mostly use preset virtual scenes that are disconnected from the user's real life and cannot establish deep memory associations. Summary of the Invention
[0003] In order to overcome the shortcomings of auxiliary memory that lacks full utilization of three-dimensional data linkage of route trajectories, real-time scene images and interpersonal relationships, the present invention provides an auxiliary vocabulary memorization method and system based on deep learning English.
[0004] The technical implementation scheme of the present invention is: an auxiliary vocabulary memorization method based on deep learning English, comprising the following steps:
[0005] S1: Obtain the user's journey data and interpersonal relationship data;
[0006] S2: Generate English vocabulary data and a first memory scene based on the user's distance data and the user's interpersonal relationship data;
[0007] S3: Obtaining a target English vocabulary memory set that the user needs to memorize, and using a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sorting the first memory scenes according to the scene recommendation value;
[0008] S4: processing the first memory scene according to the interpersonal relationship data influencing factor in the first memory scene to obtain a second memory scene;
[0009] S5: Assist the user in memorizing English vocabulary according to the second memorization scenario.
[0010] Preferably, the obtaining of the user's distance data and the user's interpersonal relationship data includes: obtaining the user's distance data and the user's interpersonal relationship data, the distance data including the user's distance trajectory and distance scene data; the interpersonal relationship data including the user's historical identification figures, the user's temporary identification figures and the user's potential identification figures; and dividing the historical identification figures into first-level historical identification figures, second-level historical identification figures and third-level historical identification figures.
[0011] Preferably, the distance data includes the user's distance trajectory and distance scene data; the interpersonal relationship data includes the user's historical identification person, the user's temporary identification person and the user's potential identification person, including: the user's distance trajectory is the walking distance of the user within the first preset time period; the user's distance scene data is the scene picture data collected after the user's portable device collects the interpersonal relationship data; the historical identification person is the identification person collected and set by the user based on the interpersonal relationship; the user's temporary identification person is the temporary person manually marked by the user; the user's potential identification person is a person whose appearance number is greater than the preset threshold and is identified by the portable device within the second preset time period.
[0012] Preferably, the English vocabulary data and the first memory scene are generated based on the user's route data and the user's interpersonal relationship data, including: when there are historical identification figures or temporary identification figures in the user's route scene data, the route scene data and the language data to be converted of the corresponding figures are recorded, the language data to be converted is converted into English vocabulary data using a deep learning model, and then the first memory scene is generated in combination with the route scene data; when there are no historical identification figures and temporary identification figures in the user's route scene data, it is detected whether there are potential identification figures, and if so, the first memory scene is generated in combination with the route scene data; if not, monitoring is continued.
[0013] Preferably, the method of obtaining the target English vocabulary memory set that the user needs to memorize, and using the scene memory recommendation formula to obtain the scene recommendation value of each first memory scene, and sorting the first memory scenes according to the scene recommendation value, includes: obtaining the target English vocabulary memory set that the user needs to memorize, obtaining the number of English vocabulary matches and matching English vocabulary based on the converted English vocabulary data and the target English vocabulary memory set, and obtaining the user's English vocabulary proficiency based on the matching English vocabulary, using the scene memory recommendation formula to obtain the scene recommendation value of each first memory scene based on the user's English vocabulary proficiency, the number of English vocabulary matches and the user's interpersonal relationship data, and sorting the first memory scenes according to the scene recommendation value.
[0014] Preferably, the scene recommendation value of each first memory scene is obtained using a scene memory recommendation formula based on the user's English vocabulary proficiency, the number of English vocabulary matches and the user's interpersonal relationship data, including: wherein the scene memory recommendation formula is:
[0015]
[0016] Where P is the scene recommendation value; n is the number of English vocabulary matches; x j is the proficiency of the jth English vocabulary; R i is the influencing factor of the interpersonal relationship data of the i-th character in the first memory scene; α, β, and ε are adjustment coefficients.
[0017] Preferably, the R i The interpersonal relationship data influence factor of the i-th character in the first memory scene is obtained according to a display influence formula, wherein the display influence formula is:
[0018]
[0019] Where R i is the interpersonal relationship data influence factor of the i-th character in the first memory scene; ω hk is the weight coefficient of the k-th level historical identification figure; C hk is the number of times the k-th level historical identification figure is identified in the first preset time period; ω t is the weight coefficient of the temporary identification character; C t is the number of times the temporary identification character is identified; ω q is the weight of potential identification person; C q The number of times a potential person is identified.
[0020] Preferably, the first memory scene is processed according to the interpersonal relationship data influencing factor in the first memory scene to obtain the second memory scene, including: processing the first memory scene after being sorted according to the scene recommendation value to obtain the second memory scene, matching the preset geographical or scene-related vocabulary library according to the geographical location information in the user's route trajectory and the visual features of the route scene data, generating background English vocabulary to be displayed in the first memory scene, and selecting the identification person with the largest interpersonal relationship data influencing factor to display the English vocabulary data separately, and generating a scene guidance interface for the remaining identification persons to obtain the second memory scene, the scene guidance interface is used to respond to the user's triggering operation on the scene guidance interface, and generate a derivative memory scene with the remaining identification persons as the core, wherein the remaining identification persons are identification persons with non-largest interpersonal relationship data influencing factors.
[0021] Preferably, the generating of the derived memory scene with the remaining identified person as the core includes: calling the second memory scene with the interpersonal relationship data influence factor of the remaining identified person as the largest as the derived memory scene.
[0022] Preferably, an auxiliary vocabulary memory system based on deep learning English also includes:
[0023] A data acquisition module is used to obtain the user's journey data and the user's interpersonal relationship data;
[0024] A memory scene generation module, configured to generate English vocabulary data and a first memory scene based on the user's journey data and the user's interpersonal relationship data;
[0025] A scene sorting module is used to obtain a target English vocabulary memory set that the user needs to memorize, and use a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sort the first memory scenes according to the scene recommendation value;
[0026] An influence factor acquisition module, used to obtain the influence factor of the interpersonal relationship data of each character in the first memory scene according to the display influence formula;
[0027] A memory scene processing module, configured to process the first memory scene to obtain a second memory scene based on an influencing factor of interpersonal relationship data in the first memory scene;
[0028] The derived scene calling module is used to call the second memory scene with the interpersonal relationship data influence factor of the remaining identified characters as the largest.
[0029] The beneficial effects are:
[0030] 1. This invention uses the user's real journey trajectory and collected scene images to construct a memory scene that is highly consistent with the real environment, so that vocabulary learning can be integrated into daily life;
[0031] 2. The present invention generates conversation context by combining historical, temporary and potential identifiers, and enhances memory stickiness through the social association of familiar characters;
[0032] 3. Automatically recommend optimal memory scenarios based on vocabulary matching number, proficiency, and interpersonal relationship data influencing factors to address users' weaknesses;
[0033] 4. Generate derivative scenes with the remaining characters as the core through the scene guidance interface, expand the memory network and avoid repeated learning fatigue. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 This is a flow chart of an auxiliary vocabulary memorization method based on deep learning English in the present invention;
[0035] Figure 2 This is a structural schematic diagram of an auxiliary vocabulary memory system based on deep learning English in the present invention. DETAILED DESCRIPTION
[0036] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0037] Example 1: A method for assisting vocabulary memorization based on deep learning English, such as Figure 1 As shown, the following steps are included:
[0038] S1: Obtain the user's journey data and interpersonal relationship data;
[0039] Obtain the user's route data and the user's interpersonal relationship data, wherein the route data includes the user's route trajectory and route scene data; the interpersonal relationship data includes the user's historical identifiers, the user's temporary identifiers, and the user's potential identifiers; and divide the historical identifiers into first-level historical identifiers, second-level historical identifiers, and third-level historical identifiers.
[0040] It should be explained that the geographic positioning function and video image acquisition module in the portable device (such as a smart phone, smart watch and translation glasses) are first called to automatically obtain the user's walking route information in real life within a first preset time period to form the user's journey trajectory. The journey trajectory includes structured data of GPS coordinate point sequence, timestamp and positioning accuracy, and a trajectory line graph is constructed in chronological order for subsequent association of scene images; combined with the image data collected by the camera, the key areas where the user has stayed in the above-mentioned journey trajectory are identified, and the image content in the area (including environmental scenes, facial images of characters and speech content of corresponding characters) is extracted, and the data is stored as the user's journey scene data; three types of character labels are preset, namely history and time. Historical identifiers, temporary identifiers and potential identifiers. Historical identifiers are regular social relationships actively maintained and defined by users, including relatives, colleagues, classmates and close friends. Historical identifiers are further divided into three subcategories: first-level historical identifiers, second-level historical identifiers and third-level historical identifiers. The specific division is based on the intimacy level, communication frequency, number of interactions or social tags set by the user. For example, first-level historical identifiers are user-set characters with "frequent contact and close relationship", such as spouse, parents and children; second-level historical identifiers are objects with "periodic interaction and medium social frequency", such as colleagues and roommates; third-level historical identifiers are characters with "occasionally contact, tags set but low interaction frequency", such as friends met at an event.
[0041] The user's route trajectory is the walking distance of the user within the first preset time period; the user's route scene data is the scene picture data collected after the user's portable device collects interpersonal relationship data; the historical identification characters are identification characters collected and set by the user based on interpersonal relationships; the user's temporary identification characters are temporary characters manually marked by the user; the user's potential identification characters are characters identified by the portable device within the second preset time period and whose appearance times are greater than a preset threshold.
[0042] It should be explained that the user's temporarily identified person is a temporary person manually marked by the user, and is a temporary social object actively marked by the user in a specific scene. For example, in a group photo temporarily taken by the user in a party, course or short-term project, the person is manually marked by clicking on the face. This type of person does not have a long-term record and is only used for memory task generation within the current or subsequent limited time period; the user's potentially identified person is a person whose number of appearances is greater than a preset threshold identified by the portable device within the second preset time period. For example, the second preset time period is set to the last 30 days. When the frequency of appearance of a face that has not been marked by the user in the route scene data exceeds the system-set threshold (such as greater than or equal to 5 times), the person will be automatically identified as a potentially identified person. The potentially identified person is dynamically generated and has strong adaptability to supplement people that the user has not actively set but often contacts in reality.
[0043] S2: Generate English vocabulary data and a first memory scene based on the user's distance data and the user's interpersonal relationship data;
[0044] When there are historical or temporary identification figures in the user's route scene data, the route scene data and the language data to be converted of the corresponding figures are recorded, and the language data to be converted is converted into English vocabulary data using a deep learning model, and then the first memory scene is generated in combination with the route scene data; when there are no historical or temporary identification figures in the user's route scene data, it is detected whether there is a potential identification figure, and if so, the first memory scene is generated in combination with the route scene data; if not, monitoring continues.
[0045] It should be explained that, first, the collected route scene data is subjected to image analysis and face recognition to determine whether historical identification figures and temporary identification figures appear in the scene picture. For example, the scene picture is input into the face detection module for comparison with the user's face library. If the first-level historical identification figure Zhang San previously set by the user and the temporary identification figure Li Si manually marked by the user in this party are matched, it is determined that the conditions of this step are met. When the above two types of identification figures are detected, the language data to be converted of the scene is automatically recorded immediately. The language data to be converted includes environmental text and character speech. The environmental text and character speech are input into the language conversion model trained by the multimodal Transformer to obtain the corresponding English vocabulary data. For example, the environmental text "Welcome to the library" is converted into "Welcome-to-the-library" or the character speech "Today "Let's go to the library together" is converted into "Let's-go-to-the-library-together-today". Based on the converted English vocabulary data and the original route scene picture, the first memory scene is synthesized according to the preset scene template, which specifically includes: a scene picture thumbnail; the converted English sentence and the corresponding Chinese interpretation; the "identified person portrait" and "person label" (historical or temporary) that triggers the scene; if no historical identified person or temporary identified person is detected, the potential identified person detection process is entered to detect whether there is a potential identified person in the route scene picture (an unlabeled person whose number of appearances exceeds a threshold within a second preset time period). If so, the first memory scene is generated according to the same process, and the person type is marked as a potential identified person; if any identified person conditions are still not met, the scene data is continuously monitored until a picture that meets the conditions is detected.
[0046] S3: Obtaining a target English vocabulary memory set that the user needs to memorize, and using a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sorting the first memory scenes according to the scene recommendation value;
[0047] Obtain a target English vocabulary memory set that the user needs to memorize, obtain the number of English vocabulary matches and matching English vocabulary based on the converted English vocabulary data and the target English vocabulary memory set, obtain the user's English vocabulary proficiency based on the matching English vocabulary, use a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene based on the user's English vocabulary proficiency, the number of English vocabulary matches and the user's interpersonal relationship data, and sort the first memory scenes according to the scene recommendation value.
[0048] It needs to be explained that the target English vocabulary memory set is loaded from the user learning plan module. This set can be manually input by the user or preset by the course outline. For each generated first memory scene, its English vocabulary data list is extracted and cross-compared with the target English vocabulary memory set to count the number of English vocabulary matches; the user vocabulary test module or learning log is called to obtain the English vocabulary proficiency of each matching vocabulary; after completing the above-mentioned scenario recommendation value calculation for all first memory scenes, they are sorted from large to small according to the scenario recommendation value to generate the final first memory scene recommendation list.
[0049] The recommended formula for scene memory is:
[0050]
[0051] Where P is the scene recommendation value; n is the number of English vocabulary matches; x j is the proficiency of the jth English vocabulary; R i is the influencing factor of the interpersonal relationship data of the i-th character in the first memory scene; α, β, and ε are adjustment coefficients.
[0052] It should be explained that P is the scenario recommendation value. The higher the value, the higher the scenario priority. ln(n+1) is the logarithm of the number of English vocabulary matches, indicating the diminishing returns effect. i R i It is the sum of interpersonal influences, reflecting the collective importance of the characters in the scene; Prioritize review and learning for low-proficiency vocabulary scenarios.
[0053] The interpersonal relationship data influence factor of the i-th character in the first memory scene is obtained according to the display influence formula, where the display influence formula is:
[0054]
[0055] Where R i is the interpersonal relationship data influence factor of the i-th character in the first memory scene; ω hk is the weight coefficient of the k-th level historical identification figure; C hk is the number of times the k-th level historical identification figure is identified in the first preset time period; ω t is the weight coefficient of the temporary identification character; C t is the number of times the temporary identification character is identified; ω q is the weight of potential identification person; C q The number of times a potential person is identified.
[0056] It should be explained that historical identifiers are divided into three levels (k = 1, 2, 3), allowing different importance (such as the first level has a higher weight); temporary and potential identifiers are calculated separately; the interpersonal relationship data influence factors of other characters in the same scene are calculated separately, and all interpersonal relationship data influence factors are used as ∑ i R i input.
[0057] S4: processing the first memory scene according to the interpersonal relationship data influencing factor in the first memory scene to obtain a second memory scene;
[0058] S5: Assist the user in memorizing English vocabulary according to the second memorization scenario.
[0059] The first memory scene sorted according to the scene recommendation value is processed to obtain the second memory scene, and the preset geographical or scene-related vocabulary library is matched according to the geographical location information in the user's route trajectory and the visual features of the route scene data to generate background English vocabulary for display in the first memory scene, and the identification person with the largest interpersonal relationship data influence factor is selected to display the English vocabulary data separately, and at the same time, a scene guidance interface is generated for the remaining identification persons to obtain the second memory scene, and the scene guidance interface is used to respond to the user's triggering operation on the scene guidance interface to generate a derivative memory scene with the remaining identification persons as the core, wherein the remaining identification persons are identification persons with non-largest interpersonal relationship data influence factors.
[0060] It should be explained that, based on the final first memory scene recommendation list, several scenes with high rankings are selected as objects to be processed. For a certain sorted first memory scene, the geographic location information in its route trajectory (such as the GPS coordinates of a coffee shop in the city center) and the visual features of the route scene data (such as the coffee cup, bar and barista that appear in the scene picture) are first extracted. Then, the preset geographic or scene-related vocabulary library is called. The vocabulary library has mapped common geographic landmarks and environmental elements to corresponding English vocabulary lists (such as "coffee-cup", "barista" and "espresso-machine"). Background English vocabulary that is highly matched with the current scene is retrieved from the library and used as background display content and embedded in the first memory scene interface. From the set of interpersonal relationship data influencing factors of all the identified characters in the first memory scene, identify the identified character corresponding to the maximum value, and for the core character, display its corresponding English vocabulary data (such as "Let's-enjoy-our-coffee-together" converted from its speech) separately in the form of bubbles, so that users can pay attention to the core character-related memory content at a glance, and generate corresponding scene guidance interface buttons or trigger areas for the remaining identified characters (such as Li Si and Wang Wu), and embed them into the current first memory scene interface to generate a second memory scene. When the user triggers the guidance interface of a remaining identified character, he can enter the next step of the derivative memory scene generation process to achieve deep memory of the character's speech content.
[0061] The second memory scene with the largest influence factor of the interpersonal relationship data of the remaining identified characters is called as the derived memory scene.
[0062] It needs to be explained that the second memory scene with the remaining identified character as the core character is selected as the derivative memory scene and displayed.
[0063] Example 2: Based on Example 1, an auxiliary vocabulary memory system based on deep learning English, such as Figure 2 As shown, it also includes:
[0064] A data acquisition module is used to obtain the user's journey data and the user's interpersonal relationship data;
[0065] A memory scene generation module, configured to generate English vocabulary data and a first memory scene based on the user's journey data and the user's interpersonal relationship data;
[0066] A scene sorting module is used to obtain a target English vocabulary memory set that the user needs to memorize, and use a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sort the first memory scenes according to the scene recommendation value;
[0067] An influence factor acquisition module, used to obtain the influence factor of the interpersonal relationship data of each character in the first memory scene according to the display influence formula;
[0068] A memory scene processing module, configured to process the first memory scene to obtain a second memory scene based on an influencing factor of interpersonal relationship data in the first memory scene;
[0069] The derived scene calling module is used to call the second memory scene with the interpersonal relationship data influence factor of the remaining identified characters as the largest.
[0070] The above is a detailed introduction to the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of the present application. At the same time, for those skilled in the art, based on the idea of the present application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.
Claims
1. An auxiliary vocabulary memorization method based on deep learning English, characterized in that: The following steps are involved: S1: Obtain the user's journey data and interpersonal relationship data; S2: Generate English vocabulary data and a first memory scene based on the user's distance data and the user's interpersonal relationship data; S3: Obtaining a target English vocabulary memory set that the user needs to memorize, and using a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sorting the first memory scenes according to the scene recommendation value; S4: processing the first memory scene according to the interpersonal relationship data influencing factor in the first memory scene to obtain a second memory scene; S5: Assist the user in memorizing English vocabulary according to the second memorization scenario.
2. The auxiliary vocabulary memorization method based on deep learning English according to claim 1 is characterized in that: The obtaining of the user's route data and the user's interpersonal relationship data includes: obtaining the user's route data and the user's interpersonal relationship data, wherein the route data includes the user's route trajectory and route scene data; the interpersonal relationship data includes the user's historical identifiers, the user's temporary identifiers, and the user's potential identifiers; and dividing the historical identifiers into first-level historical identifiers, second-level historical identifiers, and third-level historical identifiers.
3. The auxiliary vocabulary memorization method based on deep learning English according to claim 2 is characterized in that: The distance data includes the user's distance trajectory and distance scene data; the interpersonal relationship data includes the user's historical identification person, the user's temporary identification person and the user's potential identification person, including: the user's distance trajectory is the walking distance of the user within the first preset time period; the user's distance scene data is the scene picture data collected after the user's portable device collects the interpersonal relationship data; the historical identification person is the identification person collected and set by the user based on the interpersonal relationship; the user's temporary identification person is the temporary person manually marked by the user; the user's potential identification person is a person whose appearance number is greater than the preset threshold and is identified by the portable device within the second preset time period.
4. The auxiliary vocabulary memorization method based on deep learning English according to claim 1 is characterized in that: The method of generating English vocabulary data and a first memory scene based on the user's route data and the user's interpersonal relationship data includes: when there are historical identification figures or temporary identification figures in the user's route scene data, recording the route scene data and the language data to be converted of the corresponding figures, converting the language data to be converted into English vocabulary data using a deep learning model, and then generating the first memory scene in combination with the route scene data; when there are no historical identification figures or temporary identification figures in the user's route scene data, detecting whether there are potential identification figures, and if so, generating the first memory scene in combination with the route scene data; if not, continuing monitoring.
5. The auxiliary vocabulary memorization method based on deep learning English according to claim 1 is characterized in that: The method of obtaining a target English vocabulary memory set that the user needs to memorize, obtaining a scene recommendation value for each first memory scene using a scene memory recommendation formula, and sorting the first memory scenes according to the scene recommendation value includes: obtaining a target English vocabulary memory set that the user needs to memorize, obtaining an English vocabulary matching number and matching English vocabulary based on the converted English vocabulary data and the target English vocabulary memory set, obtaining the user's English vocabulary proficiency based on the matching English vocabulary, obtaining a scene recommendation value for each first memory scene using a scene memory recommendation formula based on the user's English vocabulary proficiency, the English vocabulary matching number, and the user's interpersonal relationship data, and sorting the first memory scenes according to the scene recommendation value.
6. The auxiliary vocabulary memorization method based on deep learning English according to claim 5, characterized in that: The method of obtaining a scene recommendation value for each first memory scene using a scene memory recommendation formula based on the user's English vocabulary proficiency, the number of English vocabulary matches, and the user's interpersonal relationship data includes: The recommended formula for scene memory is: Where P is the scene recommendation value; n is the number of English vocabulary matches; x j is the proficiency of the jth English vocabulary; R i is the influencing factor of the interpersonal relationship data of the i-th character in the first memory scene; α, β, and ε are adjustment coefficients.
7. The auxiliary vocabulary memorization method based on deep learning English according to claim 6, characterized in that: The R i is the influencing factor of the interpersonal relationship data of the i-th character in the first memory scene, including: The interpersonal relationship data influence factor of the i-th character in the first memory scene is obtained according to the display influence formula, where the display influence formula is: Where R i is the interpersonal relationship data influence factor of the i-th character in the first memory scene; ω hk is the weight coefficient of the k-th level historical identification figure; C hk is the number of times the k-th level historical identification figure is identified in the first preset time period; ω t is the weight coefficient of the temporary identification character; C t The number of times a person is temporarily identified; ω q is the weight of potential identification person; C q The number of times a potential person is identified.
8. The auxiliary vocabulary memorization method based on deep learning English according to claim 1 is characterized in that: The method of processing the first memory scene according to the interpersonal relationship data influencing factor in the first memory scene to obtain the second memory scene includes: processing the first memory scene after being sorted according to the scene recommendation value to obtain the second memory scene, matching a preset geographical or scene-related vocabulary library according to the geographical location information in the user's route trajectory and the visual features of the route scene data, generating background English vocabulary to be displayed in the first memory scene, and selecting the identification person with the largest interpersonal relationship data influencing factor to display the English vocabulary data separately, and generating a scene guidance interface for the remaining identification persons to obtain the second memory scene, and the scene guidance interface is used to respond to the user's triggering operation on the scene guidance interface to generate a derivative memory scene with the remaining identification persons as the core, wherein the remaining identification persons are identification persons with non-largest interpersonal relationship data influencing factors.
9. The auxiliary vocabulary memorization method based on deep learning English according to claim 8, characterized in that: The generating of the derived memory scene with the remaining identified person as the core includes: calling a second memory scene with the interpersonal relationship data influence factor of the remaining identified person as the largest as the derived memory scene.
10. An auxiliary vocabulary memory system based on deep learning English, according to the auxiliary vocabulary memory method based on deep learning English according to any one of claims 1 to 9, characterized in that: Also includes: A data acquisition module is used to obtain the user's journey data and the user's interpersonal relationship data; A memory scene generation module, configured to generate English vocabulary data and a first memory scene based on the user's journey data and the user's interpersonal relationship data; A scene sorting module is used to obtain a target English vocabulary memory set that the user needs to memorize, and use a scene memory recommendation formula to obtain a scene recommendation value for each first memory scene, and sort the first memory scenes according to the scene recommendation value; An influence factor acquisition module, used to obtain the influence factor of the interpersonal relationship data of each character in the first memory scene according to the display influence formula; A memory scene processing module, configured to process the first memory scene to obtain a second memory scene based on an influencing factor of interpersonal relationship data in the first memory scene; The derived scene calling module is used to call the second memory scene with the interpersonal relationship data influence factor of the remaining identified characters as the largest.