Photographer screening optimization method for photography appointment shooting platform

Abstract

The invention discloses a photographer screening optimization method for a photography appointment platform, and the method comprises the following steps: S1, inputting a screening condition by a user, and enabling the screening condition to comprise a precondition and a shooting style; s2, generating a recommendation image set, and taking the recommendation image set as a screening sample image; s3, carrying out screening by using the sample image, and feeding back screening information items; s4, establishing a contact channel to which the entry belongs; s5, screening and optimizing; performing screening again, and repeating the step S2 and the step S3; the information items comprise pictures meeting screening conditions, photographer information, photographer gender, cooperation evaluation, price intervals, geographic positions and style labels. According to the method, the problem that the definition of the photographic sub-styles is fuzzy when only keywords and labels are adopted for searching is solved, the searching accuracy is effectively improved, and the personalized requirements of users can be fully met.

Description

Technical Field

[0001] The present invention relates to the technical field of information retrieval, and specifically to an optimization method for screening photographers on a photography appointment platform. Background Art

[0002] With the popularization of social media and mobile devices, users' demand for personalized photography services is increasing continuously. Traditional photography appointment platforms mainly rely on simple search and sorting methods. Users need to manually screen through a large number of photographers, shooting themes, and styles. This is time-consuming and laborious, and there is also a certain fuzzy interval in the definition between photography styles. There are subjective differences in the definition of a certain shooting sample style, which leads to the situation that some users cannot find completely matching tags for the shooting styles they need on the photography platform, and thus cannot meet the shooting requirements. Users need to spend a lot of time screening and conducting complex tag or keyword searches.

[0003] Therefore, the existing photography appointment screening methods have problems such as low efficiency, insufficient personalization, and low search accuracy. The key lies in that the retrieval methods using style tags and keywords are prone to ignoring the subtle differences in user needs, resulting in inaccurate recommendation results. Based on the problems existing in the above-mentioned photography appointment screening process, the present invention proposes an optimization method for screening photographers on a photography appointment platform. Summary of the Invention

[0004] The purpose of the present invention is to provide an optimization method for screening photographers on a photography appointment platform to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An optimization method for screening photographers on a photography appointment platform, including the following steps:

[0006] S1: The user inputs screening conditions, and the screening conditions include preconditions and shooting styles;

[0007] S2: Generate a recommended picture set and use it as a screening sample picture;

[0008] S3: Screen with the sample picture and feedback the screening information entries;

[0009] S4: Construct a contact channel for the photographers to whom the entries belong;

[0010] S5: Screening optimization; conduct screening again and repeat steps S2 and S3.

[0011] Preferably, the information entries include pictures that meet the screening conditions, photographer information, photographer gender, cooperation evaluation, price range, geographical location, and style tags.

[0012] Preferably, in step S1:

[0013] The pre-screening conditions include price range, gender of the photographer, shooting theme, distance of the photographer, and level of the photographer;

[0014] The shooting style selection includes manual selection and importing pictures for recognition;

[0015] The range of manual style selection is the style tags already recorded in the photography appointment platform. Users can manually select tags and adjust the proportion of the selected tags. The features for importing pictures for recognition are based on the style tags in the manual selection, and a tag set composed of different style tags is generated;

[0016] The picture recognition also includes the features of scenery in the picture, tone, texture, light, perspective, and emotional elements, generating a feature tag set corresponding to the imported picture. The style tag set and the feature tag set form an identification element tag collection.

[0017] Preferably, in step S2:

[0018] Based on the identification element tag collection obtained by importing pictures for recognition in step S1, the tag information in the identification element tag collection is read. In the way of AI picture generation, the elements in the identification element tag collection are embodied, generating sample pictures that conform to the identification element tag collection. The number of generated sample pictures is 1 - 10;

[0019] The generated sample picture set is fed back to the user terminal for the user to screen. The user selects the sample pictures that meet the requirements as the basis for subsequent screening;

[0020] If there are no sample pictures that meet the user's requirements, step S5 is executed.

[0021] Preferably, in step S3:

[0022] According to the generated sample picture set selected in step S2, the database extracts the types of tags and the generation ratio between each tag in the generated sample picture set as a reference template, and uses it as the data set for screening;

[0023] The data set is matched with the information entries in the database. Referring to the types of sample picture tags and the ratio between tags in the database, a matching coincidence degree threshold is set. The threshold range is 30 - 100%. The sample picture information entries with the coincidence degree within the threshold range are added to the candidate list, and the display order of the candidate list is proportional to the coincidence degree. The higher the coincidence degree, the higher the position of the information entry in the candidate list;

[0024] After the candidate list is sorted, in combination with the pre-screening conditions in step S1, match the price range, photographer's gender, shooting theme, photographer's distance, and photographer's level of the information entries, and perform a secondary sorting based on the degree of matching. The entries that more conform to the user's matching conditions will be displayed higher, obtaining a comprehensive screening information entry list for the user to browse and select.

[0025] Preferably, in step S4:

[0026] If there are information entries that meet the user's requirements, establish a connection channel between the photographer to whom the information entry belongs and the user;

[0027] If there are no information entries that meet the user's requirements, perform step S5 for screening and optimization.

[0028] Preferably, in step S5:

[0029] The screening and optimization include manual optimization and automatic optimization;

[0030] The manual optimization includes adjusting the selection range of the preconditions, adjusting the selection style tags, style element ratios, re-importing picture recognition, and increasing or decreasing the number of imported recognition pictures;

[0031] The automatic optimization includes the appointment shooting platform identifying the user account's browsing album records, extracting the tags of the album with the highest browsing frequency and the longest browsing time, generating a tag data set, and integrating the tag data set into the recognition element tag collection in step S1;

[0032] After screening and optimization, generate sample pictures again by AI. The user selects the sample pictures as the screening template, and after screening, feedback the screening information entries for the user to select.

[0033] Preferably, the manual optimization also includes adjusting the ratios of the various tags in the recognition element tag collection. The recognition element tag collection after adjusting the tag ratios generates a new data set and performs AI to generate sample pictures.

[0034] A photography appointment shooting platform for a photographer screening and optimization method, characterized in that: the photography appointment shooting platform includes a client, a platform terminal, a photographer terminal, and a database, and a network connection is established among the client, the platform terminal, the photographer terminal, and the database using the Internet;

[0035] Among them, the photographer terminal uploads the photographer's picture albums, photographer information, photographer's gender, price range, geographical location, and style tags. The platform terminal evaluates the professionalism and cooperation evaluation indicators based on the photographer's past cooperation quality and adjusts the price range;

[0036] The platform terminal performs label recognition and creation on the work pictures uploaded by the photographer in sequence, assigns the corresponding weight item values to the labels, and stores them in the database in a classified manner.

[0037] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0038] 1. The present invention adopts the method of generating AI sample pictures according to user needs, achieving the effect of quickly understanding user needs. By using AI image recognition to extract multiple labels and generating AI sample pictures again based on the extracted labels, various user needs can be combined to obtain sample pictures that better meet user needs. This not only meets the personalized needs of users but also solves the problem of users' vague definition of shooting styles, effectively shortening the time for users to screen.

[0039] 2. The present invention uses the method of taking the AI-generated sample picture as a screening template to further meet the personalized needs of users. Taking the AI-generated sample picture selected by the user as a screening template, while meeting the basic shooting needs of users, the user demand labels are spread to the entire database, and information entries with a matching coincidence degree are matched from the database, effectively improving the screening accuracy and obtaining photographic information that meets the personalized needs of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of the screening optimization process in the first embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0042] Please refer to Figure 1 , the present invention provides the following two embodiments:

[0043] Embodiment 1:

[0044] Please refer to Figure 1 , a method for optimizing the screening of photographers on a photography appointment platform, including the following steps:

[0045] S1: The user inputs screening conditions, and the screening conditions include preconditions and shooting styles.

[0046] The screening preconditions include price range, photographer's gender, shooting theme, photographer's distance, and photographer's level. The price range, photographer's gender, shooting theme, photographer's distance, and photographer's level all contain sub-menus. Among them, the price range sub-menu includes 100 - 500, 500 - 700, 700 - 800, etc.; the photographer's gender sub-menu includes male, female, and no limit; the shooting theme sub-menu includes wedding, personal portrait, baby photo, one-year-old photo, etc.; the photographer's distance sub-menu includes within 1 km, within 2 km, within 5 km, within 10 km, etc.; the photographer's level sub-menu includes level one to level five.

[0047] The shooting style selection includes manual selection and importing pictures for recognition.

[0048] The range of manually selected styles is the style tags already recorded in the photography appointment platform. Users can manually select tags and adjust the proportions of the selected tags. The tags of manually selected styles are the style categories generally recognized by the public, and the total proportion of manually selected style tags is 1.

[0049] The features for importing pictures for recognition are based on the style tags in manual selection, generating a tag set composed of different style tags.

[0050] Picture recognition also includes features of scenery in the picture, color tone, texture, light, perspective, and emotional elements, etc. Among them, the scenery features include trees, lawns, mountains, flowing water, the sea, etc.; the color tone features include warm color tone, cold color tone, black and white color tone, etc.; the texture features include oil painting, realistic, blurred, etc.; the perspective features include wide-angle, aerial photography, close-up, etc.; the emotional features include joy, cheerfulness, liveliness, etc. Generate a feature tag set corresponding to the imported picture. The main body of the feature tag set is the percentage data set of recognition features. The style tag set and the feature tag set form an identification element tag collection.

[0051] Among them, after the data sets of the feature tag set and the style tag set are merged, the total proportion of each feature and tag is 1, that is, the proportion of features and tags is expressed in percentage form, and a data set is generated.

[0052] S2: Generate a recommended picture set and use it as a screening sample picture.

[0053] Based on the identification element tag collection obtained from importing pictures for recognition in step S1, read the tag information in the identification element tag collection, and in the way of AI picture generation, reflect the elements in the identification element tag collection to generate sample pictures that conform to the identification element tag collection. The number of generated sample pictures is 1 - 10.

[0054] Generate a set of sample images and feedback them to the user terminal for the user to screen. The user selects the sample images that meet the requirements as the basis for subsequent screening. Extract multiple tags through AI image recognition, and use the extracted tags to generate sample images by AI again. The multiple requirements of the user can be combined to obtain sample images that better meet the user's needs, which not only meets the personalized needs of the user, but also solves the problem of the user's vague definition of the shooting style, effectively shortening the user's screening time.

[0055] For AI image generation, DALL-E drawing is adopted to extract the elements in the recognition element tag set. For example, the features extracted from the recognition element tag set are: wedding, outdoor, joy, warm color tone, sea, wide angle. Then a wedding dress sample image taken by the sea in a wide-angle manner is generated as a reference for subsequent screening. When optimizing the screening, the extracted features can be displayed as a percentage, which can be directly adjusted to control the proportion of the features for screening optimization.

[0056] If there are no sample images that meet the user's requirements, execute step S5 to directly perform generation optimization.

[0057] S3: Screen with the sample images and feedback the screening information entries.

[0058] The information entries include the images that meet the screening conditions, photographer information, photographer gender, cooperation evaluation, price range, geographical location, and style tags.

[0059] According to the generated set of sample images selected in step S2, the database uses the generated set of sample images as a reference template to extract the types of tags in the generated set of sample images and the generation ratio between the tags as the data set for screening.

[0060] Match the data set with the information entries in the database. Refer to the types of sample image tags and the ratio between the tags in the database, and set the matching overlap threshold. The threshold range is 30 - 100%. The sample image information entries with the overlap within the threshold range are added to the candidate list, and the display order of the candidate list is proportional to the degree of overlap. The higher the degree of overlap, the higher the position of the information entry in the candidate list.

[0061] The overlap calculation includes preprocessing, feature extraction, and similarity calculation. Among them, preprocessing uniformly processes the images, such as size adjustment, color, and spatial transformation, to ensure the standardization of the calculation.

[0062] Feature extraction uses convolutional neural network (CNN) feature extraction. A pre-trained deep learning model is used to extract the high-level features of the image, and the cosine similarity or Euclidean distance between the feature vectors is calculated as the similarity of the image, which can capture complex image patterns and composition elements.

[0063] Calculate the similarity between features, generate a similarity matrix or scores, and perform screening and classification according to the overlap degree by setting a threshold.

[0064] After the candidate list is arranged, in combination with the pre-screening conditions in step S1, match the price range, photographer's gender, shooting theme, photographer's distance, and photographer's level of the information entries, and perform a secondary sorting according to the matching degree. The entries that more conform to the user's matching conditions will be displayed higher, and a comprehensive screening information entry list will be obtained for the user to browse and select.

[0065] Using the AI-generated sample image selected by the user as a screening template, while meeting the user's basic shooting needs, spread the user's demand tags to the entire database, and match the information entries with a coincidence degree that meets the requirements from the database, effectively improving the screening accuracy and obtaining photographic information that meets the user's personalized needs.

[0066] S4: Construct a contact channel for the photographer to whom the entry belongs.

[0067] If there are information entries that meet the user's requirements, then build a contact channel between the photographer to whom the information entry belongs and the user.

[0068] If there are no information entries that meet the user's requirements, then perform step S5 for screening optimization.

[0069] S5: Screening optimization; perform screening again, repeating step S2 and step S3.

[0070] The screening optimization includes manual optimization and automatic optimization.

[0071] The manual optimization includes adjusting the selection range of preconditions, adjusting the selected style tags, style element ratios, re-importing pictures for recognition, and increasing or decreasing the number of imported pictures for recognition.

[0072] The automatic optimization includes the appointment platform identifying the user account's browsing record of picture sets, extracting the tags of the picture set with the highest browsing frequency and the longest browsing time, generating a tag data set, and integrating the tag data set into the recognition element tag collection in step S1.

[0073] After screening optimization, generate an AI-generated sample image again. The user selects the sample image as a screening template, and after screening, feedback the screened information entries for the user to select.

[0074] The manual optimization also includes adjusting the ratio of each tag in the recognition element tag collection. After adjusting the tag ratio, a new data set is generated from the recognition element tag collection, and an AI-generated sample image is performed.

[0075] Set the feature matrix and accuracy score for the fused feature vector set, construct a random forest model, and use the test set to evaluate the model performance. Calculate metrics such as mean squared error. After training the random forest model, extract the feature importance. The higher the feature importance, the more significant the impact of the feature on the prediction result. Convert the feature importance into a weight vector and perform normalization processing. Apply the obtained weight vector to the user's original feature data to obtain the adjusted feature vector. Finally, use the adjusted feature vector and combine it with the similarity calculation algorithm to generate the final recommendation result.

[0076] By fusing the weight terms in the feature vector set, optimized screening conditions can be obtained, which is conducive to recommending information that better meets the personalized needs of users, further reducing the time for users to screen information, and obtaining photography information that better meets their own needs.

[0077] Embodiment 2:

[0078] The photography appointment platform includes a client, a platform terminal, a photographer terminal, and a database. The client, the platform terminal, the photographer terminal, and the database are connected to form a network connection via the Internet.

[0079] Among them, the photographer terminal uploads the photographer's picture collection, photographer information, photographer's gender, professionalism, price range, geographical location, and style tags. The platform terminal evaluates the professionalism and cooperation evaluation indicators based on the previous cooperation quality of the photographer and adjusts the price range.

[0080] The platform terminal performs label recognition and creation on the work pictures uploaded by the photographer in sequence, assigns the corresponding weight term values to the labels, and stores them in the database in a classified manner.

[0081] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A photographer screening optimization method for a photography booking platform, characterized by: The steps include: S1: The user inputs the filtering conditions, which include the preconditions and the shooting style; S2: Generate recommended atlases and use them as screening samples; S3: Screening with sample pictures and feedback of screening information items; S4: Establish a contact channel for photographers of the entries; S5: Screening optimization; perform screening again and repeat steps S2 and S3.

2. The method for selecting and optimizing photographers for a photography booking platform according to claim 1, characterized in that: Information items include images that meet the filtering criteria, photographer information, photographer gender, collaboration reviews, price range, geographic location and style tags.

3. The method for optimizing the selection of photographers for a photography booking platform according to claim 2, characterized in that: in, In step S1: Pre-screening conditions included price range, photographer gender, subject matter, photographer distance, and photographer level; Shooting style selection includes manual selection and imported picture recognition; The manually selected style range is the style tags that have been recorded in the photography booking platform. Users can manually select tags and adjust the proportion of selected tags. The image recognition features are imported based on the manually selected style tags to generate a tag set composed of different style tags. Image recognition also includes the characteristics of the scenery, color, texture, light, perspective and emotional elements in the image, generating a set of feature labels corresponding to the imported image. The style label set and the feature label set constitute the collection of identification element labels.

4. The method for selecting and optimizing photographers for a photography booking platform according to claim 3, characterized in that: in, In step S2: Based on the identification element label collection obtained by importing image recognition in step S1, the label information in the identification element label collection is read, and the elements in the identification element label collection are reflected in the form of AI image generation to generate sample images that conform to the identification element label collection, and the number of generated sample images is 1-10; Generate a sample picture set and feed it back to the user terminal for the user to screen. The user selects the sample pictures that meet the requirements as the basis for subsequent screening; If there is no sample image that meets the user's requirements, step S5 is executed.

5. The method for optimizing the selection of photographers for a photography booking platform according to claim 4, characterized in that: in, In step S3: According to the generated sample atlas selected in step S2, the database uses the generated sample atlas as a reference template to extract the types of labels in the generated sample atlas and the generation ratios between the labels as the data set used for screening; Match the dataset with the information items in the database, refer to the sample image label types and the ratio between labels in the database, set the matching overlap threshold, the threshold range is 30-100%, add the sample image information items with overlap within the threshold range to the candidate list, and the display order of the candidate list is proportional to the overlap degree. The higher the overlap degree, the higher the position of the information item in the candidate list; After the candidate list is arranged, the price range, photographer gender, shooting theme, photographer distance and photographer level of the information item are matched in combination with the pre-filtering conditions in step S1, and secondary sorting is performed according to the degree of matching. The items that meet the user's matching conditions are displayed at a higher position, and a comprehensive filtered information item list is obtained for the user to browse and select.

6. The method for optimizing the selection of photographers for a photography booking platform according to claim 5, characterized in that: in, In step S4: If there is an information item that meets the user's requirements, a communication channel between the photographer of the information item and the user will be established; If there is no information item that meets the user's requirements, the process goes to step S5 to perform screening optimization.

7. The method for optimizing the selection of photographers for a photography booking platform according to claim 6, characterized in that: in, In step S5: Screening optimization includes manual optimization and automatic optimization; Manual optimization includes adjusting the selection range of preconditions, adjusting the style labels, the proportion of style elements, re-importing images for recognition, and increasing or decreasing the number of images imported for recognition; Automatic optimization includes the photo booking platform identifying the user account browsing album records, extracting the labels of the albums with the highest browsing frequency and the longest browsing time, generating a label data set, and integrating the label data set into the identification element label collection in step S1; After screening and optimization, AI generates sample images again. The user selects the sample images as screening templates. After screening, the screening information items are fed back for the user to choose.

8. The method for optimizing the selection of photographers for a photography booking platform according to claim 7, characterized in that: Manual optimization also includes adjusting the ratio of each label in the identification element label collection, generating a new data set from the identification element label collection after adjusting the label ratio, and performing AI to generate sample images.

9. A photography appointment platform for a photographer screening optimization method according to any one of claims 1 to 8, characterized in that: The photography booking platform includes a client, a platform terminal, a photographer terminal and a database, and the client, the platform terminal, the photographer terminal and the database are connected by the Internet; The photographer uploads the photographer's photo album, photographer information, photographer gender, price range, geographic location, and style tags. The platform terminal assesses the photographer's professionalism and cooperation evaluation indicators based on the quality of past cooperation and adjusts the price range. The platform terminal identifies and creates labels for the works in turn based on the collection of works uploaded by the photographer, assigns weight item values ​​corresponding to the labels, and stores them in the database by category.

Images