Artificial intelligence based data analysis method and system

Abstract

The application relates to an artificial intelligence-based data analysis method and system, and relates to the technical field of data analysis, which comprises the following steps: collecting a detection period and detection information of the detection period; updating the detection information through a preset checking method, and performing large-to-small sorting on the detection information to obtain sorting information; sequentially performing feature extraction from large to small from the sorting information to obtain feature information; calling marked feature information according to the feature information and the detection period; and generating and displaying output of detection clothes through the marked feature information. The application has the effect of improving the efficiency of data analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, and in particular to an artificial intelligence-based data analysis method and system. BACKGROUND

[0002] Data analysis is a process of extracting valuable information, discovering rules, solving problems and supporting decision-making from a large amount of data based on data collection, cleaning, conversion, modeling and interpretation.

[0003] When researching clothing data, it is usually necessary to manually research different clothing data and data existence periods from various websites, transcribe the researched data into various tables, and then arrange and integrate the data according to the required data to form a complete table. The data in the table is analyzed to obtain data results.

[0004] When researching clothing data, traditional data research and analysis rely on manual work, which is time-consuming and laborious, and reduces the efficiency of data analysis. SUMMARY

[0005] In order to improve the efficiency of data analysis, the present application provides an artificial intelligence-based data analysis method and system.

[0006] In a first aspect, the present application provides an artificial intelligence-based data analysis method, which adopts the following technical solution:

[0007] An artificial intelligence-based data analysis method comprises:

[0008] S10: Collecting detection periods and detection information of the detection periods;

[0009] S11: Updating the detection information by a preset verification method, and sorting the detection information from large to small to obtain sorted information;

[0010] S12: Sequentially extracting features from the sorted information from large to small to obtain feature information;

[0011] S13: Retrieving labeled feature information according to the feature information and the detection period;

[0012] S14: Generating and displaying output of the detection clothing through the labeled feature information.

[0013] By adopting the above technical solution, the detection information of each detection period is verified by a verification method, and the verified detection information is sorted to obtain sorted information. The feature information is extracted from the sorted information to obtain feature information, and the feature information and the detection period are analyzed to generate and display output of the detection clothing. Thus, the clothing data can be automatically collected and analyzed, and the efficiency of data analysis is improved.

[0014] Optionally, the verification method comprises:

[0015] S20: collecting detection information and detection operation information, the detection operation information comprising current operation information and historical operation information;

[0016] S21: obtaining tolerance information from the historical operation information;

[0017] S22: obtaining deviation information according to a comparison between the current operation information and the historical operation information;

[0018] S23: obtaining abnormal detection information based on an exceeding of the deviation information and the tolerance information;

[0019] S24: updating the detection information according to the abnormal detection information.

[0020] Optionally, the method of obtaining the deviation information comprises:

[0021] S30: retrieving a collection path and historical collection information from the historical operation information;

[0022] S31: retrieving detection collection information according to a containing of the current operation information and the collection path;

[0023] S32: combining the detection collection information and the detection information to obtain collection features;

[0024] S33: retrieving a page update rate from the detection collection information;

[0025] S34: collecting a page image of the page update rate;

[0026] S35: identifying page detection features from the page image;

[0027] S36: retrieving a marked update rate according to a consistency between the page detection features and the collection features;

[0028] S37: retrieving a historical update rate from the historical operation information based on the page detection features and the collection features;

[0029] S38: obtaining the deviation information according to a consistency between the marked update rate and the historical update rate.

[0030] Optionally, the method further comprises:

[0031] S40: collecting page selection information according to a consistency between the page detection features and the collection features;

[0032] S41: identifying page selection costumes based on the page selection information and the page image;

[0033] S42: identifying a selection feature of the page selecting clothes;

[0034] S43: obtaining a marked difference feature according to a comparison between the selection feature and the collection feature;

[0035] S44: obtaining a difference query path according to the page image and the marked difference feature;

[0036] S45: obtaining a historical difference path by combining the historical operation information and the difference query path;

[0037] S46: obtaining deviation information by comparing the difference query path and the historical difference path;

[0038] Optionally, the verification method of the deviation information further comprises:

[0039] S50: obtaining a historical page image from the historical difference path by comparing the difference query path and the historical difference path;

[0040] S51: obtaining historical change information and a historical change image by using the historical page image and the historical operation information;

[0041] S52: identifying a historical change feature based on the historical change image and the historical page image;

[0042] S53: obtaining detection change information corresponding to the historical change information from the current operation information;

[0043] S54: obtaining a detection change image according to the detection change information and the page image;

[0044] S55: identifying a detection change feature from the detection change image;

[0045] S56: obtaining deviation information according to a comparison between the detection change feature and the collection feature;

[0046] Optionally, the verification method of the detection information comprises:

[0047] S60: obtaining a detection image and a detection gallery of the detection information based on an exceeding condition of the deviation information and allowable deviation information;

[0048] S61: identifying interactive text from the current operation information to generate interactive semantics;

[0049] S62: searching for a marked image from the detection gallery by combining the interactive semantics and the detection image;

[0050] S63: obtaining a difference feature by comparing features extracted from the detection image and the marked image;

[0051] S64: generating an inquiry picture based on the difference feature, the detection picture and the marked picture, outputting the inquiry picture for interaction, and collecting reply information;

[0052] S65: generating a reference reply semantic through the inquiry picture;

[0053] S66: generating a reply semantic according to the reply information, and continuously updating the detection information according to a comparison between the reply semantic and the reference reply semantic.

[0054] By using the above technical solution, the inquiry picture generated by the user output through the difference feature, the detection picture and the marked picture is collected, the reply semantic corresponding to the reply information is collected, the detection information is updated according to the comparison between the reply semantic and the reference reply semantic, and the accuracy of the detection information is improved.

[0055] Optionally, the generation method of the inquiry picture comprises:

[0056] S70: collecting a color difference type and a feature difference type;

[0057] S71: obtaining a marked difference type based on the consistency of the difference feature, the color difference type and the feature difference type;

[0058] S72: identifying a detection clothing color from the detection picture;

[0059] S73: generating a sticker text according to the detection clothing color;

[0060] S74: updating the detection picture and the marked picture through the sticker text;

[0061] S75: combining the detection picture and the marked picture to obtain the inquiry picture.

[0062] Optionally, the verification method of the detection picture and the marked picture further comprises:

[0063] S80: collecting a position difference type and a contour difference type;

[0064] S81: obtaining a target difference type based on the consistency of the difference feature, the color difference type and the feature difference type;

[0065] S82: obtaining target difference information through the target difference type and the difference feature;

[0066] S83: obtaining selected difference information according to the consistency of the target difference information, the position difference type and the contour difference type;

[0067] S84: identifying a detection clothing posture through the detection picture;

[0068] S85: obtaining detection difference information according to the detection clothing posture and the difference feature;

[0069] S86: obtaining marked difference information based on the detection clothing posture and the difference feature and the marked picture;

[0070] S87: selecting a clothing posture from the detection clothing postures based on consistency of the detection difference information and the marked difference information;

[0071] S88: updating the marked picture and the detection picture according to the selected clothing posture.

[0072] Optionally, the method further comprises:

[0073] S90: obtaining selected difference information according to consistency of the target difference information, the position difference type and the contour difference type;

[0074] S91: identifying a detection feature contour of the difference feature from the detection picture;

[0075] S92: identifying a marked feature contour of the difference feature from the marked picture;

[0076] S93: comparing the marked feature contour and the detection feature contour to form an overlapping contour;

[0077] S94: analyzing the overlapping contour, the detection picture and the marked picture to obtain a change wrinkle;

[0078] S95: updating the marked picture and the detection picture according to the change wrinkle.

[0079] By using the above technical solution, the marked picture and the detection picture are updated by analyzing the color and the contour of the type of the difference feature, and the updated marked picture and the detection picture are combined to obtain the inquiry picture, so that the accuracy of identifying the user operation can be improved, and the accuracy of the detection information can be further improved.

[0080] In a second aspect, the present application provides a data analysis system based on artificial intelligence, which adopts the following technical solution:

[0081] A data analysis system based on artificial intelligence, comprising:

[0082] An acquisition module for acquiring detection information and a detection period;

[0083] A memory for storing a program of a data analysis method based on artificial intelligence;

[0084] A processor for loading and executing the program stored in the memory.

[0085] In summary, the present application includes at least one of the following beneficial technical effects:

[0086] 1. The detection information of each detection period is checked by a checking method, and the checked detection information is sorted to obtain sorting information, the sorting information is feature extracted to obtain feature information, and the feature information and the detection period are analyzed to generate and display output the detection clothes, so that the clothes data can be automatically collected and analyzed, and the efficiency of data analysis is improved;

[0087] 2. The inquiry picture generated by the difference feature, the detection picture and the marked picture is output to the user, and the reply information corresponding to the reply semantics is collected, and the detection information is updated according to the comparison between the reply semantics and the reference reply semantics, so that the accuracy of the detection information is improved;

[0088] 3. The type of the difference feature is analyzed to update the marked picture and the detection picture, and the updated marked picture and the detection picture are combined to obtain the inquiry picture, so that the accuracy of identifying the user operation is improved, and the accuracy of the detection information is further improved. BRIEF DESCRIPTION OF DRAWINGS

[0089] Figure 1 is a method flow chart of a data analysis method based on artificial intelligence according to an embodiment of the application;

[0090] Figure 2 is a method flow chart of a detection information checking method according to an embodiment of the application;

[0091] Figure 3 is a method flow chart of an inquiry picture generation method according to an embodiment of the application. DETAILED DESCRIPTION

[0092] The application will be further described in detail below with reference to the drawings and embodiments.

[0093] Referring to Figure 1 , the application discloses a data analysis method based on artificial intelligence, comprising the following steps:

[0094] S10: collecting a detection period and detection information of the detection period.

[0095] The detection period refers to a time period for which data collection is required, and the detection information refers to data information corresponding to the detection period. The detection period and the detection information are obtained after being pre-input by an operator. In this embodiment, the detection information can be the sales volume of clothes in different periods.

[0096] S11: updating the detection information by a preset checking method, and sorting the detection information from large to small to obtain sorting information.

[0097] The verification method is a method set by the technician for verifying whether the detection information is accurate. After the verification method, new detection information can be obtained.

[0098] The sorting information refers to the detection information after sorting. The total information integrated by sorting the detection information from large to small is the sorting information.

[0099] S12: Feature extraction is performed from large to small in the sorting information to obtain feature information.

[0100] The feature information refers to the key features that can reflect the detection information extracted from the sorting information. By extracting features from each detection information in the sorting information from large to small, until covering a preset number (such as the top 50 of sales ranking) or meeting a preset threshold (such as the sales of the top few clothing in sales ranking, accounting for 70% of all sales), the feature information corresponding to the detection information is obtained. The extraction method of feature information is known to those skilled in the art, and will not be repeated here.

[0101] In this embodiment, the feature information is the fabric category, color combination, pattern, and style type of the clothing.

[0102] S13: According to the feature information and the detection period, the marked feature information is retrieved.

[0103] The marked feature information refers to the feature information selected in combination with the detection period. By comparing the feature information corresponding to the detection information in the upward trend between each detection period, the marked feature information is obtained. For example, the feature information corresponding to the detection information of 70 in May is 80 in June, and at this time, the feature information is the marked feature information.

[0104] S14: Generate and display output the detection clothing through the marked feature information.

[0105] The clothing library is a database set by the technician for recording different clothing. The clothing library is updated in real time. The display area is an area set by the technician for displaying clothing.

[0106] The detection clothing refers to the clothing that does not appear in the clothing library. By combining and analyzing the marked feature information, the detection clothing is obtained, and the picture corresponding to the detection clothing is output to the display area. The generation method of the detection clothing is known to those skilled in the art, and will not be repeated here.

[0107] The verification method includes:

[0108] S20: Collect the detection device and detection operation information of the detection information. The detection operation information includes current operation information and historical operation information.

[0109] The detection device refers to a hardware device for transmitting detection information, and the detection device includes a user terminal and a server, etc. The detection operation information refers to all operation behavior information related to the detection information, the current operation information refers to the detection operation information in the current period, and the historical operation information refers to the detection operation information before the current period. The detection device and the detection operation information can be obtained by querying the system for the device corresponding to the detection information. The detection operation information includes browsing, adding to the shopping cart, placing an order, browsing time, page refresh times, and clicking and sliding paths of the page corresponding to the detection information, etc.

[0110] S21: Obtain the allowable deviation information through the historical operation information.

[0111] The allowable deviation information refers to the normal deviation range allowed for the user to operate. The allowable deviation information is obtained by analyzing the historical operation information. For example, the page refresh times in the historical operation information are 3 to 5 times, the browsing time of a commodity is within 20 minutes to 30 minutes, the page sliding speed changes at 3 pictures or 5 pictures per second, etc. These information are taken as the allowable deviation information.

[0112] S22: Obtain the deviation information according to the comparison between the current operation information and the historical operation information.

[0113] The deviation information refers to the actual difference information between the current operation information and the historical operation information. The information generated by comparing and analyzing the corresponding data of the current operation information and the historical operation information is taken as the deviation information.

[0114] S23: Obtain the abnormal detection information based on the exceeding situation of the deviation information and the allowable deviation information.

[0115] The abnormal detection information refers to the detection information corresponding to the deviation information exceeding the allowable deviation information. The deviation information and the allowable deviation information are analyzed to determine whether the deviation information exceeds the allowable deviation information. When the deviation information exceeds the allowable deviation information, it means that the detection information corresponding to the deviation information is not output by the user. The detection information corresponding to the deviation information exceeding the allowable deviation information is taken as the abnormal detection information.

[0116] When the deviation information does not exceed the allowable deviation information, it means that the detection information corresponding to the deviation information is output by the user, and no adjustment is made.

[0117] S24: Update the detection information according to the abnormal detection information.

[0118] The abnormal detection information in each detection information is removed to obtain new detection information. In this embodiment, the new detection information is the update of the sales quantity of the clothing, and the clothing corresponding to the detection information is not removed.

[0119] The method for obtaining the deviation information comprises:

[0120] S30: retrieve the collection path and the historical collection information from the historical operation information.

[0121] The collection path refers to the complete operation steps of the user for clothing collection, and the historical collection information refers to the complete operation steps of the user for clothing collection before the current period. The collection path and the historical collection information are retrieved from the historical operation information.

[0122] S31: retrieve the detection collection information according to the inclusion of the current operation information and the collection path.

[0123] The detection collection information refers to the operation step information corresponding to the collection path in the current operation information. When the current operation information includes the collection path, the operation information corresponding to the collection path in the current operation information is taken as the detection collection information.

[0124] S32: combine the detection collection information and the detection information to obtain the collection feature.

[0125] The collection feature refers to the clothing feature currently collected by the user. The feature information of the clothing corresponding to the detection collection information in the detection information is taken as the collection feature.

[0126] S33: retrieve the page update rate from the detection collection information.

[0127] The page update rate refers to the content update times of the product page corresponding to the detection collection information in unit time. The page update rate is retrieved from the detection collection information.

[0128] S34: collect the page image of the page update rate.

[0129] The page image refers to the visual image obtained by regularly taking screenshots of the product page within the page update rate statistical period. The image taken by the system within the page update rate statistical period can be taken as the page image.

[0130] S35: identify the page detection feature from the page image.

[0131] The page detection feature refers to the clothing feature extracted from the page image. The page detection feature is identified from the page image.

[0132] S36: retrieve the marked update rate by the consistency of the page detection feature and the collection feature.

[0133] The label update rate refers to the page update rate of the consistent page detection feature and the collection feature corresponding to the time period. The page update rate corresponding to the time period of the page detection feature consistent with the collection feature is called, and the called page update rate is taken as the label update rate.

[0134] S37: The historical update rate is called from the historical operation information based on the page detection feature and the collection feature.

[0135] The historical update rate refers to the page update rate of the consistent page detection feature and the collection feature corresponding to the time period before the current time period. The historical update rate is called from the historical operation information according to S36. The calling method of the historical update rate is known to those skilled in the art, and will not be repeated here.

[0136] S38: Obtain the deviation information according to the consistency of the label update rate and the historical update rate.

[0137] By analyzing the consistency of the label update rate and the historical update rate, when the label update rate and the historical update rate are inconsistent, the difference between the label update rate and the historical update rate is calculated as the deviation information.

[0138] Also includes:

[0139] S40: Collect page selection information through the consistency of the page detection feature and the collection feature.

[0140] The page selection information refers to the information of the page corresponding to the consistent page detection feature and the collection feature. The operation information of the time period corresponding to the page detection feature consistent with the collection feature is collected by the system as the page selection information.

[0141] S41: Identify page selection clothing based on page selection information and page image.

[0142] The page selection clothing refers to the selected clothing in the page image. The page selection clothing is obtained by analyzing the page selection information and the page image. The analysis method of the page selection clothing is known to those skilled in the art, and will not be repeated here.

[0143] S42: Identify the selection feature of the page selection clothing.

[0144] The selection feature refers to the feature of the page selection clothing, which is identified from the page selection clothing.

[0145] S43: Obtain the label difference feature according to the comparison of the selection feature and the collection feature.

[0146] The marked difference feature refers to a feature that appears different between the selected feature and the collection feature. The feature that appears different after comparison between the selected feature and the collection feature is taken as the marked difference feature.

[0147] S44: obtaining a difference query path according to the page image and the marked difference feature.

[0148] The difference query path refers to a query operation path performed by the user to solve the marked difference feature. The difference query path is obtained by analyzing the change of the page image and the marked difference feature. For example, for color difference, the user selects the detail page of clothing from the page, filters the color interval, and views the clothing in the updated page image. The above path is the difference query path.

[0149] S45: obtaining a historical difference path by combining the historical operation information and the difference query path.

[0150] The historical difference path refers to a query path in the historical operation information that is consistent with the difference query path. The query path in the historical operation information that is consistent with the difference query path is taken as the historical difference path with reference to S44.

[0151] S46: obtaining deviation information by comparing the consistency of the difference query path and the historical difference path.

[0152] By comparing the consistency of the difference query path and the historical difference path, when the difference query path and the historical difference path are inconsistent, the matching degree (such as the step coincidence rate and the operation time difference) of the difference query path and the historical difference path is calculated. If the matching degree is lower than a preset threshold (fifty percent) or the operation time difference exceeds the normal range (such as the current path time is three times the historical path time), the difference value is the deviation information.

[0153] The verification method of the deviation information further includes:

[0154] S50: comparing the consistency of the difference query path and the historical difference path to retrieve a historical page image from the historical difference path.

[0155] The historical page image refers to a page image corresponding to the historical difference path that is consistent with the difference query path. The page image corresponding to the path that is consistent with the difference query path is retrieved from the historical difference path as the historical page image.

[0156] S51: obtaining historical change information and a historical change image by using the historical page image and the historical operation information.

[0157] The historical change information refers to operation information of changes in page content of the historical page image caused by user operation, and the operation information that changes the historical page image is called as the historical change information.

[0158] The historical change image refers to a page changed by operation of the historical change information, and a page image changed in a time period of the historical change information is called as the historical change image. In this embodiment, the historical change image is an enlarged historical page image, and the historical change information is operation of triggering the historical change image to be enlarged.

[0159] S52: Identify the historical change feature based on the historical change image and the historical page image.

[0160] The historical change feature refers to a feature of changed clothes in the historical change image, and a feature of clothes changed in the historical change image is called as the historical change feature by comparing the historical change image with the historical page image.

[0161] S53: Call detection change information corresponding to the historical change information from the current operation information.

[0162] The detection change information refers to operation information of changed features of the page image in the current operation information, and the detection change information corresponding to the historical change information is called from the current operation information.

[0163] S54: Obtain a detection change image according to the detection change information and the page image.

[0164] The detection change image refers to a page image changed by operation of the detection change information, and the detection change image is obtained by analyzing the detection change information and the page image. The analysis method of the detection change image is well known to those skilled in the art, and will not be described here.

[0165] S55: Identify the detection change feature from the detection change image.

[0166] The detection change feature refers to a feature of changed clothes in the detection change image, and the changed clothes feature in the detection change image is called as the detection change feature by referring to S52.

[0167] S56: Obtain deviation information according to consistency of the detection change feature and the collection feature.

[0168] The deviation information is obtained by analyzing the difference between the detection change feature and the collection feature. For example, the difference between the fabric and the pattern, the difference degree is matched from the preset clothes comparison table by the detection change feature and the collection feature, and the parameter value corresponding to the difference degree is called as the deviation information.

[0169] The clothing comparison table stores different detection change features and collection features corresponding to the degree of difference, such as the degree of difference between the fabric and the pattern, the degree of difference between the color and the fabric, and the like. The parameters in the clothing comparison table are set by the person skilled in the art in advance according to actual conditions, and are not described herein.

[0170] With reference to Figure 2 The verification method of the detection information includes the following steps.

[0171] S60: Based on the exceeding condition of the deviation information and the allowable deviation information, the detection picture of the detection information and the detection gallery are collected.

[0172] The detection picture refers to the clothing picture output by the user for communication in the detection information, and the detection gallery refers to the database of the clothing collected by the system. By analyzing the exceeding condition of the deviation information and the allowable deviation information, when the deviation information is less than the allowable deviation information, it indicates that there is still a situation not operated by the user himself, and then the clothing picture output by the user for communication in the system during the detection information period and the detection gallery are retrieved, and the output clothing picture is taken as the detection picture.

[0173] S61: The interactive text is recognized from the current operation information to generate the interactive semantics.

[0174] The interactive text refers to the text output by the user for communication in the detection information, and the interactive semantics refers to the semantics of the interactive text. The interactive text is recognized from the current operation information, and the interactive semantics is generated by analyzing the interactive text. The recognition and generation method of the interactive text and the interactive semantics is well known to the person skilled in the art, and is not described herein.

[0175] S62: The marked picture is searched from the detection gallery in combination with the interactive semantics and the detection picture.

[0176] The marked picture refers to the picture in the detection gallery which matches the interactive semantics and the features of the detection picture. The marked picture corresponding to the features is searched from the detection gallery by analyzing the total features of the clothing obtained by combining the interactive semantics and the detection picture. The searching and analyzing method of the marked picture is well known to the person skilled in the art, and is not described herein.

[0177] S63: The features extracted between the detection picture and the marked picture are compared to obtain the difference features.

[0178] The difference features refer to the clothing features that appear different between the detection picture and the marked picture. The features extracted between the detection picture and the marked picture are compared, and the inconsistent features after comparison are taken as the difference features.

[0179] S64: Based on the difference features, the detection picture and the marked picture, the inquiry picture is generated and outputted for interaction, and the reply information is collected.

[0180] The inquiry picture refers to a picture used for communication inquiry with the user, and the reply information refers to information replied by the user when receiving the inquiry picture. The inquiry picture is obtained by analyzing the difference feature, the detection picture and the marked picture, and is output to an area for communication with the user to interact with the user, and then the text output by the user within a preset time period (for example, 12 hours) after the inquiry picture is called by the system as the reply information.

[0181] S65: generating a reference reply semantic by the inquiry picture.

[0182] The reference reply semantic refers to a semantic normally replied after the inquiry picture is output, and is obtained by analyzing the inquiry picture. The analysis method of the reference reply semantic is known to those skilled in the art, and is not described here.

[0183] S66: generating a reply semantic according to the reply information, and continuing to update the detection information according to the comparison between the reply semantic and the reference reply semantic.

[0184] The reply semantic refers to a semantic expressed by the reply information, and is obtained by analyzing the text of the reply information. Then, the reply semantic is compared with the reference reply semantic, and the detection information corresponding to the reply semantic inconsistent with the reference reply semantic is excluded to obtain new detection information with reference to S24.

[0185] With reference to Figure 3 , the generation method of the inquiry picture comprises:

[0186] S70: collecting color difference types and feature difference types.

[0187] The color difference type refers to a type of color difference, and the feature difference type refers to a type of feature difference. The color difference type and the feature difference type are obtained by pre-input of an operator.

[0188] S71: obtaining a marked difference type based on the consistency of the difference feature, the color difference type and the feature difference type.

[0189] The marked difference type refers to a type of difference feature consistent with the color difference type. The type of difference feature, the color difference type and the feature difference type are compared, and when the type of difference feature is consistent with the color difference type, the type of difference feature is taken as the marked difference type.

[0190] S72: identifying a detection clothing color from the detection picture.

[0191] The detection clothing color refers to a color of clothing in the detection picture, and is identified from the detection picture.

[0192] S73: generating a map text according to the detected clothing color.

[0193] The map text refers to a text with a semantic of detecting the clothing color, and the map text is generated by analyzing the detected clothing color.

[0194] S74: updating the detected picture and the marked picture by the map text.

[0195] The map text is combined in the detected picture and the marked picture to generate new detected picture and marked picture.

[0196] S75: combining the detected picture and the marked picture to obtain an inquiry picture.

[0197] The inquiry picture is obtained by combining the detected picture and the marked picture. For example, the left side of the inquiry picture is the detected picture, and the right side is the marked picture. The left and right positions of the detected picture and the marked picture are randomly adjusted, and the reference reply semantic changes correspondingly with the left and right positions of the detected picture and the marked picture.

[0198] The verification method of the detected picture and the marked picture further comprises:

[0199] S80: collecting a position difference type and a contour difference type.

[0200] The position difference type refers to a type for describing the position difference of the clothing feature, and the contour difference type refers to a type for describing the contour difference of the clothing feature, which is obtained by pre-inputting by an operator.

[0201] S81: obtaining a target difference type based on the consistency of the difference feature, the color difference type and the feature difference type.

[0202] The target difference type refers to the type of the difference feature consistent with the feature difference type, which is obtained by comparing the type of the difference feature, the color difference type and the feature difference type, and when the type of the difference feature is consistent with the feature difference type, the type of the difference feature is taken as the target difference type.

[0203] S82: obtaining target difference information by the target difference type and the difference feature.

[0204] The target difference information refers to the position and shape information of the difference feature of the target difference type, which is obtained by analyzing the target difference type and the difference feature.

[0205] S83: obtaining selected difference information according to the consistency of the target difference information, the position difference type and the contour difference type.

[0206] The selected difference information refers to the type of the target difference information consistent with the position difference type. The target difference information type, the position difference type, and the contour difference type are compared. When the target difference information type is consistent with the position difference type, the target difference information of the position difference type is selected as the selected difference information.

[0207] S84: Recognize the detected clothing pose by detecting the picture.

[0208] The detected clothing pose refers to the display pose of the clothing in the detected picture. The clothing in the detected picture is analyzed to obtain the detected clothing pose. In this embodiment, the detected clothing pose is a plurality of poses.

[0209] S85: Obtain the detected difference information according to the detected clothing pose and the difference feature.

[0210] The detected difference information refers to the position and shape information of the difference feature in the detected clothing pose. The position and shape information corresponding to the difference feature in the detected clothing pose are analyzed as the detected difference information. The analysis method of the detected difference information is known to those skilled in the art, and will not be repeated here.

[0211] S86: Obtain the marked difference information based on the detected clothing pose, the difference feature, and the marked picture.

[0212] The marked difference information refers to the position and shape information of the difference feature in each detected clothing pose in the marked picture. When the clothing pose in the marked picture is consistent with the detected clothing pose, the position and shape information of the difference feature in each pose in the marked picture are recognized as the marked difference information.

[0213] S87: Retrieve the selected clothing pose from the detected clothing pose based on the consistency of the detected difference information and the marked difference information.

[0214] The selected clothing pose refers to the clothing pose whose detected difference information is consistent with the marked difference information. The detected clothing pose corresponding to the consistency of the detected difference information and the marked difference information is defined as the selected clothing pose.

[0215] S88: Update the marked picture and the detected picture according to the selected clothing pose.

[0216] The clothing pose in the marked picture and the detected picture is consistent with the selected clothing pose. The updated marked picture and the detected picture are used as new marked picture and detected picture.

[0217] Also includes:

[0218] S90: obtaining selected difference information according to consistency of target difference information, position difference type and contour difference type.

[0219] The selected difference information refers to the type of the target difference information consistent with the contour difference type, and the type of the target difference information, the position difference type and the contour difference type are compared, and when the type of the target difference information is consistent with the contour difference type, the target difference information of the contour difference type is selected as the selected difference information.

[0220] S91: identifying a detection feature contour of the difference feature from the detection picture.

[0221] The detection feature contour refers to the contour corresponding to the difference feature in the detection picture, and the contour of the difference feature identified from the detection picture is taken as the detection feature contour.

[0222] S92: identifying a mark feature contour of the difference feature from the mark picture.

[0223] The mark feature contour refers to the contour corresponding to the difference feature in the mark picture, and the contour of the difference feature identified from the mark picture is taken as the mark feature contour.

[0224] S93: comparing the mark feature contour and the detection feature contour to form an overlapping contour.

[0225] The overlapping contour refers to the contour that appears in the mark feature contour and the detection feature contour, and the range contour of the two contours that overlap is taken as the overlapping contour by comparing the mark feature contour and the detection feature contour.

[0226] S94: analyzing the overlapping contour, the detection picture and the mark picture to obtain a change wrinkle.

[0227] The change wrinkle refers to the wrinkle that needs to be changed for the contour of the difference feature in the detection picture and the mark picture to become the overlapping contour, and the detection picture and the mark picture are analyzed respectively, and the wrinkle corresponding to the contour of the difference feature of the clothes in the detection picture and the mark picture that drives the contour to change to the overlapping contour is taken as the change wrinkle. The analysis method of the change wrinkle is the common knowledge of those skilled in the art, which is not described here.

[0228] S95: updating the mark picture and the detection picture according to the change wrinkle.

[0229] The clothes in the mark picture and the detection picture are updated by the change wrinkle, and the mark picture and the detection picture with the change wrinkle are taken as new mark picture and detection picture.

[0230] Based on the same inventive concept, the embodiments of the present application provide a data analysis system based on artificial intelligence, comprising:

[0231] The acquisition module is configured to acquire a detection period, detection information, a detection device, detection operation information, a page image, page selection information, a detection picture, a detection gallery, reply information, a color difference type, a feature difference type, a position difference type, and a contour difference type.

[0232] The memory is configured to store a program of the artificial intelligence-based data analysis method.

[0233] The processor is configured to load and execute the program stored in the memory.

[0234] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional modules is exemplified, and in actual application, the above functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. The specific working process of the above-described system, device and unit can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0235] The above is only the preferred embodiment of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the concept of the present application shall be deemed to fall within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application shall also be deemed to fall within the protection scope of the present application.

Claims

1. A data analysis method based on artificial intelligence, characterized in that, include: S10: Collect the detection period and detection information for the detection period; S11: Update the detection information using a preset verification method, and sort the detection information from largest to smallest to obtain sorting information; S12: Extract features from the sorted information in descending order to obtain feature information; S13: Retrieve the labeled feature information based on the feature information and the detection time period; S14: Generate detected clothing by marking feature information and display the output; Verification methods include: S20: Detection equipment and detection operation information for collecting detection information, including current operation information and historical operation information; S21: Obtain allowable deviation information through historical operation information; S22: Obtain deviation information by comparing current operation information with historical operation information; S23: Obtain anomaly detection information based on the exceedance of deviation information and allowable deviation information; S24: Update the detection information based on the anomaly detection information; Methods for obtaining deviation information include: S30: Retrieve collection path and historical collection information from historical operation information; S31: Retrieve and detect collection information based on the current operation information and the contents of the collection path; S32: Combine the collection information with the detection information to obtain collection characteristics; S33: Retrieve page update rate from collection information; S34: Capture page images showing the page update rate; S35: Identify page detection features from the page image; S36: Determine the tag update rate by checking the consistency between page detection features and collection features; S37: Retrieve historical update rates from historical operation information based on page detection features and collection features; S38: Obtain deviation information based on the consistency between the marker update rate and the historical update rate; Also includes: S40: Collect page selection information by checking the consistency between page detection features and collection features; S41: Identify clothing selected on the page based on page selection information and page image; S42: Identify the selection characteristics of clothing on the page; S43: Based on the comparison between the selected features and the collection features, the marking difference features are obtained; S44: Obtain the difference query path based on the difference features between the page image and the marker; S45: Combine historical operation information with the difference query path to obtain the historical difference path; S46: Compare the consistency between the difference query path and the historical difference path to obtain deviation information; Methods for verifying deviation information also include: S50: Compare the consistency between the difference query path and the historical difference path, and retrieve the historical page image from the historical difference path; S51: Obtain historical change information and historical change images by using historical page images and historical operation information; S52: Identify historical change features based on historical change images and historical page images; S53: Retrieve the detection change information corresponding to the historical change information from the current operation information; S54: Obtain the detected change image based on the detected change information and the page image; S55: Identify the features of detected changes from the image of detected changes; S56: Obtain deviation information based on the consistency between the detected change characteristics and the collection characteristics.

2. The data analysis method based on artificial intelligence according to claim 1, characterized in that, The verification methods for detection information include: S60: Based on the exceedance of deviation information and allowable deviation information, collect detection images and detection image library of detection information; S61: Identify interactive text from current operation information to generate interactive semantics; S62: Combine interactive semantics with the detected images to find labeled images from the detection image library; S63: Compare the features extracted between the detected image and the labeled image to obtain the difference features; S64: Generate a query image based on differential features, detected images, and labeled images, output the query image for interaction, and collect response information; S65: Generate baseline response semantics by querying images; S66: Generate response semantics based on the response information, and continue to update the detection information based on the comparison between the response semantics and the baseline response semantics.

3. The data analysis method based on artificial intelligence according to claim 2, characterized in that, The methods for generating the images include: S70: Collect color difference types and feature difference types; S71: Obtain the labeling difference type based on the consistency between the difference features, color difference type, and feature difference type; S72: Identify and detect clothing colors from the detected image; S73: Generate textured text based on detected clothing color; S74: Update detected and labeled images using textured text; S75: Combine the detected image and the labeled image to obtain the query image.

4. The data analysis method based on artificial intelligence according to claim 3, characterized in that, The verification methods for detecting and labeling images also include: S80: Acquisition location difference type and contour difference type; S81: Obtain the target difference type based on the consistency between the difference features, color difference type, and feature difference type; S82: Obtain target difference information by identifying target difference types and difference features; S83: Obtain the selection difference information based on the consistency of target difference information, position difference type and contour difference type; S84: Detect clothing poses by examining images; S85: Obtain detection difference information based on the detected clothing posture and difference characteristics; S86: Obtain labeling difference information based on detected clothing posture and differential features, as well as labeled images; S87: Based on the consistency between the detected difference information and the labeled difference information, retrieve and select the clothing posture from the detected clothing posture; S88: Update the marked image and the detection image based on the selected clothing posture.

5. The data analysis method based on artificial intelligence according to claim 4, characterized in that, Also includes: S90: Selected difference information is obtained based on the consistency of target difference information, position difference type and contour difference type; S91: Detection feature contours for identifying differential features from detection images; S92: Marked feature contours for identifying differential features from marked images; S93: Compare the marked feature contours with the detected feature contours to form overlapping contours; S94: Analyze overlapping contours and detected images, as well as labeled images, to obtain varied wrinkles; S95: Update the labeled image and the detected image based on the changes in wrinkles.

6. A data analysis system based on artificial intelligence, characterized in that, include: The acquisition module is used to acquire detection information and detection time periods; A memory for storing a program that implements the artificial intelligence-based data analysis method as described in any one of claims 1 to 5; The processor is used to load and execute programs stored in memory.

Images