Product planning support device and product planning support method

Abstract

To provide a product planning support device and method that can predict product design plans and sales volumes based on data.SOLUTION: A product planning support device according to one embodiment includes: an interface unit for data input / output; and a product planning unit connected to the interface unit. The product planning unit classifies one or more search images searched based on at least one of one or more collected keywords into one or more groups, receives a reference image selection input to select at least one of the one or more search images classified into the one or more groups from a user through the interface unit, receives image editing prompt data from the user through the interface unit, and can output a product design image based on the one or more images selected in the reference image selection input and the image editing prompt data.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a technology for supporting product planning, and in particular to a product planning support device and a product planning support method that can predict product design plans and sales volumes based on data. [Background technology]

[0002] In recent years, image editing technology using artificial intelligence models has been actively researched, and product design development methods have been developed using this technology. In particular, various models have been developed that edit input images in response to specific text prompts.

[0003] Furthermore, various methods for forecasting sales volume using a forecasting model have been attempted. However, most forecasting models predict future sales volume based on past sales volume, and there is a problem that the accuracy of the forecast value is significantly reduced if there is an error in the past data or if unexpected variables different from past sales occur.

[0004] Patent Document 1 discloses a technology relating to a trend forecast-based product plan generation method and device. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Korean Patent No. 10-2609682 Summary of the Invention [Problem to be solved by the invention]

[0006] The object is to provide a product planning support device and a product planning support method that can predict product design plans and sales volumes based on data. [Means for solving the problem]

[0007] According to one aspect, a product planning support device includes an interface unit for data input / output; and a product planning unit connected to the interface unit, wherein the product planning unit classifies one or more search images searched based on at least one of one or more collected keywords into one or more groups, receives a reference image selection input from a user through the interface unit to select at least one of the one or more search images classified into the one or more groups, receives image editing prompt data from the user through the interface unit, and outputs a product design image based on the one or more images selected in the reference image selection input and the image editing prompt data.

[0008] The product planning department classifies the collected one or more keywords into major category keywords and subcategory keywords according to predetermined rules, receives a user's keyword selection input for at least one of the major category keywords and subcategory keywords through the interface department, and can output search volume by date based on the user's keyword selection input.

[0009] The product planning unit may receive input for at least one of a search platform, a search keyword, and a search period through the interface unit, search for images from the search platform based on the search keyword and the search period to generate one or more search images, and classify the one or more search images to generate one or more groups.

[0010] The product planning department may receive an input regarding the number of crowds from a user, determine a predetermined classification criterion according to the number of crowds, and generate crowds of the number of crowds based on the predetermined classification criterion determined according to the number of crowds.

[0011] The product planning department can estimate the expected sales volume based on the forecast input data including at least one of the product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price.

[0012] The forecasted sales volume may be sales volume data by date.

[0013] The product planning department can search for one or more similar products based on the predicted input data.

[0014] The product planning department can extract comparison input data corresponding to the forecast input data for one or more similar products, and estimate the expected sales volume based on the comparison input data for each of the one or more similar products.

[0015] The product planning department may retrieve and output actual sales data for one or more similar products along with estimated expected sales volumes for each of the one or more similar products.

[0016] The product planning department can output comparison input data corresponding to the predicted input data for one or more similar products.

[0017] According to one aspect, a product planning support method performed on a computing device having one or more processors and memory storing one or more programs executed by the one or more processors includes the steps of: classifying one or more search images searched based on at least one of one or more collected keywords into one or more groups; receiving a reference image selection input from a user to select at least one of the one or more search images classified into one or more groups; receiving image editing prompt data from the user; and outputting a product design image based on the one or more images selected in the reference image selection input and the image editing prompt data. [Effects of the Invention]

[0018] By providing a data-based product design planning device, the time and cost of product design can be dramatically reduced.

[0019] In addition, sales volume can be predicted more accurately through sales forecasting and searching for similar products, and fluctuations in sales volume due to external variables can be intuitively understood. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a configuration diagram of a product planning support device according to an embodiment; [Figure 2] 1 is an exemplary diagram for explaining a product design planning method according to an example; [Figure 3] 1 is an exemplary diagram for explaining a product design planning method according to an example; [Figure 4] 1 is an exemplary diagram for explaining a product design planning method according to an example; [Figure 5] 1 is an exemplary diagram for explaining a product design planning method according to an example; [Figure 6] 1 is an exemplary diagram for explaining a product design planning method according to an example; [Figure 7] FIG. 10 is an exemplary diagram illustrating a method for predicting product sales volume according to an example. [Figure 8] FIG. 10 is an exemplary diagram illustrating a method for predicting product sales volume according to an example. [Figure 9] 1 is a flowchart illustrating a product planning support method according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0021] Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings. When describing the present invention, detailed descriptions of related known functions or configurations are omitted if it is determined that such descriptions may unnecessarily obscure the gist of the present invention. Furthermore, the terms used below are defined in consideration of the functions of the present invention, and may vary depending on the intentions or practices of users or operators. Therefore, the definitions should be based on the overall content of this specification.

[0022] Hereinafter, an embodiment of a product planning support device and method will be described in detail with reference to the drawings.

[0023] FIG. 1 is a configuration diagram of a product planning support device according to an embodiment.

[0024] Referring to FIG. 1, a product planning support device 100 may include an interface unit 110 for data input / output and a product planning unit 120 connected to the interface unit 110 .

[0025] According to one embodiment of the present invention, the product planning support device 100 can collect at least one of data on clothing design images, clothing appearance images, keywords for clothing types, sales volume, and user reviews from various web / app pages, such as online shopping malls selling clothing and in-house malls of design companies. To this end, for example, the product planning support device 100 can store and execute a crawling agent. The crawling agent can be considered a software module that can, for example, search multiple web pages on the Internet and analyze the content of those web pages. The crawling agent can be used to collect the various data described above.

[0026] In one embodiment, the product planning support device 100 can convert a clothing design into text or generate a clothing design based on the text. For example, the clothing design can be various text data that describes characteristics of the clothing, such as material, size, shape, and gender. Furthermore, since the text data describes the characteristics of the clothing, the clothing design can be generated based on the text data.

[0027] In another embodiment, the product planning support device 100 can generate a new clothing design based on the input of a clothing design and text. For example, a new clothing design can be generated based on the input clothing design and text, for example, on data input related to user preferences, fashion data, etc.

[0028] To implement the above embodiment, the product planning support device 100 may form a network with at least one generative artificial neural network. Specifically, the product planning support device 100 may construct a pre-trained clothing design-text-based multimodal design generation model and generate clothing designs using the model. The clothing design-text-based multimodal design generation model is a model that can learn by combining (associating) and processing text, images, videos, audio, etc., into which multiple clothing designs are converted.

[0029] In this way, the product planning support device 100 can generate new clothing design images using embedding vectors for text and / or image input data. Embedding vectors are vectorized by converting text and / or image input data into numerical values or numerical sequences that indicate their meanings and relationships. New clothing design images can be generated using these embedding vectors. In particular, according to an embodiment of the present invention, the product planning support device 100 can select multiple images generated by initial input as reference image data, and generate a final design image by inputting the reference image data or the reference image data and image editing prompts.

[0030] For example, the interface unit 110 may include an interface module such as a monitor, keyboard, mouse, touch screen, etc. for exchanging data with a user. For another example, the interface unit 110 may include a communication module for performing data input / output with an external server, user terminal, etc.

[0031] According to one embodiment, the product planning unit 120 may classify one or more search images searched based on at least one of the collected keywords into one or more groups. A search image is an image obtained as a result of a search based on the collected keywords and is an image expressed by the collected keywords. For example, the image may represent the overall appearance of an object, including the type, shape, color, etc., of the object of the product planning.

[0032] According to one embodiment, the product planning unit 120 classifies one or more collected keywords into major category keywords and subcategory keywords according to predetermined rules, receives a user's keyword selection input for at least one of the major category keywords and subcategory keywords through the interface unit 110, and can output search volume by date based on the user's keyword selection input.

[0033] For example, the product planning department 120 may collect keywords according to a predetermined criterion from an external server. For example, the product planning department 120 may calculate the search volume by time period for the collected keywords or determine the keyword rankings based on the search volume.

[0034] Referring to Fig. 2, the product planning department 120 can classify keywords by predetermined period, such as the last week, the last month, or the last three months. In this case, the product planning department 120 can determine a ranking based on the search volume for each period. Referring to Fig. 3, the product planning department 120 can output a change in search volume by date for each keyword.

[0035] As an example, the product planning department 120 may group similar keywords to categorize them. For example, the product planning department 120 may categorize blue jeans and jean as blue jeans. The grouping of similar keywords is not limited to this example, but the product planning department 120 may have a table or the like indicating the similarity of keywords in advance, and may group keywords based on this similarity. Furthermore, the table or the like may be updated through learning. Thereafter, the product planning department 120 may calculate the search volume of the categorized keywords using the search volume of the grouped similar keywords.

[0036] According to one embodiment, the product planning unit 120 receives input for at least one of a search platform, a search keyword, and a search period through the interface unit 110, searches for images from the search platform based on the search keyword and the search period to generate one or more search images, and classifies the one or more search images to generate one or more groups.

[0037] According to one example, the product planning department 120 may receive an input from a user selecting a keyword for at least one of the searched keywords and may search for images based on the input keyword. In this case, the product planning department 120 may receive an input for at least one of a search platform, a search keyword, and a search period, and may search for images from a search platform based on the input search keyword and search period to generate one or more search images. For example, the product planning department 120 may receive an input of a search keyword such as "fur, down coat," and may receive input of data for a search platform and a search period such as "Platform A" and "Last year." In this case, the product planning department 120 may search for images based on the input conditions and download the searched image to generate a search image.

[0038] For example, the product planning department 120 may search for images based on at least one of image metadata and image description data stored in an external server. As another example, the product planning department 120 may include a neural network model trained to perform keyword-based image searches, and may classify images searched through the neural network model to classify the images by search keywords.

[0039] According to an example, the product planning department 120 may classify one or more search images to generate one or more groups. For example, the product planning department 120 may classify the search images according to preset classification criteria. For example, the preset classification criteria may include the type of clothing, color, material, user's gender, length, presence or absence of printing, etc., and the product planning department 120 may classify the search images according to some or a combination of the preset classification criteria.

[0040] According to one example, the product planning unit 120 may include a neural network model for performing image clustering. The neural network model may utilize a conventional image clustering model. The image clustering model is a model that clusters data based on the similarity between the data.

[0041] According to one embodiment, the product planning department 120 may receive an input regarding the number of crowds from a user and determine predetermined classification criteria based on the number of crowds. For example, the product planning department 120 may determine classification criteria by using some or combining pre-set classification criteria such as the type of clothing, color, material, user's gender, length, and whether or not there is printing, based on the number of crowds, and may classify search images based on the determined classification criteria.

[0042] According to one embodiment, the product planning unit 120 may generate crowds corresponding to the number of crowds based on predetermined classification criteria determined by the number of crowds. For example, as shown in FIG. 4(a), if a user requests 10 crowds, the product planning unit 120 may classify the search image into 10 crowds and output them. In FIG. 4(a), one point is matched to one search image, and when the user selects one point, the matched search image may be output.

[0043] 4(b), the product planning department 120 may output the images classified into groups by group. For example, if the user selects group 3, the product planning department 120 may output the search images included in group 3.

[0044] According to one embodiment, the product planning unit 120 may receive a reference image selection input from a user via the interface unit 110, the reference image selection input selecting at least one of one or more search images classified into one or more groups. For example, the user may select one of a plurality of search images associated with one group and output, and determine the selected image as the reference image. The reference image may be a search image selected by the user from a group (a group generated by classifying one or more search images) that includes the reference image. The reference image may be a search image representative of the group that includes the reference image, or a search image that the user prefers among the search images included in the group. Referring to FIG. 5, the user may select one search image as a reference image and input the selected reference image.

[0045] According to one embodiment, the product planning department 120 may receive image editing prompt data from a user through the interface unit 110. For example, as shown in FIG. 5, a user may input an image editing prompt for editing a reference image input through the interface unit 110. The image editing prompt may be composed of predetermined keywords or predetermined sentences. The image editing prompt is a prompt such as text that the user conveys to the product planning department 120 to generate a desired image. The image editing prompt may include various information elements such as the product's color, material, length, age group, and sales volume. In addition, the product planning department 120 may receive input of technical parameters for image editing.

[0046] According to one embodiment, the product planning department 120 can output a product design image based on one or more images selected by a user through a reference image selection input and image editing prompt data. Referring to Figure 6, the product planning department 120 can edit the reference image based on the input image editing prompt and output the edited one or more images.

[0047] According to one embodiment, the product planning unit 120 may estimate a predicted sales volume for a product desired by a user based on predicted input data including at least one of a product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price. The predicted input data is input data for estimating a predicted sales volume for a product desired by a user. For example, the product information may include at least one of the color, material, and length of the product. The celebrity score may be determined based on at least one of the search volume, number of followers, and number of celebrities of the celebrity. The social media score may be determined based on at least one of the number of users who uploaded information about the product, the number of followers of the uploading users, search volume, and posting volume.

[0048] For example, the product planning department 120 may receive input regarding at least one of a product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price, and may predict sales volume based on the input. That is, the product planning department 120 may predict sales volume using information regarding the product itself, sales destination, sales period, and sales price, without data regarding past sales volume. In this case, the product planning department 120 may predict sales volume for a predetermined period of time, such as a year, half year, quarter, month, or week. The product planning department 120 may also calculate expected sales volume by date.

[0049] According to one embodiment, the product planning unit 120 of the product planning support device 100 may convert various data, such as the celebrity score, distribution platform, and sales price, into embedding vectors to calculate the expected sales volume, and use the embedded vectorized values as input values to predict the expected sales volume. To this end, the product planning unit 120 may be equipped with a pre-trained demand forecasting model, and when the embedding vectorized values of the various data are input to the demand forecasting model, the expected sales volume may be output. Since the expected sales volume is output as a result of a predetermined vector input, this may also be referred to as a vector search function, as if it were a vector-value-based search result. The demand forecasting model may be generated by, for example, training various data, such as the celebrity score, distribution platform, and sales price, using machine learning. Conventional demand forecasting models include, for example, a model that predicts the future expected sales volume of a product desired by a user based on the average value of past sales volumes of the product desired by the user calculated over a period such as the past three months, and a model that predicts the expected sales volume based on past sales volumes of the product desired by the user. On the other hand, the demand forecasting model related to the present invention is characterized in that various data such as the celebrity score, distribution platform, and sales price are converted into embedding vectors, and the embedded vector values are used as new input values. Furthermore, the demand forecasting model related to the present invention is capable of predicting future expected sales volumes for any product by inputting the new input values. For example, and not limited to this, if the demand forecasting model is not a demand forecasting model that takes into account past sales volumes for a product desired by a user, but a demand forecasting model that can predict future expected sales volumes for various other products by taking into account the new input values, then by using the demand forecasting model, it may be possible to predict the expected sales volumes for a product desired by a user, even if data on past sales volumes for the product desired by the user is not available. Note that multiple demand forecasting models may be prepared, each with different types of basic data for constructing the model. The product planning department 120 can select from the multiple demand forecasting models a demand forecasting model that can perform optimal demand forecasting based on input values. For example, if one demand forecasting model is generated with a focus on celebrity scores and celebrity scores are included in the input values from users, etc., the product planning department 120 can selectively use the demand forecasting model that takes the celebrity scores into consideration to perform demand forecasting. Furthermore, if the type of data used to generate the demand forecasting model is highly correlated with the input values from users, etc., for example, if the product planning department 120 detects a high degree of agreement between the type of data and the input values, it can perform demand forecasting using the demand forecasting model.

[0050] For example, the product planning unit 120 of the product planning support device 100 can convert the results of the vector input into text or quantity, and in the above example, can convert the resulting vector value into a predicted sales volume and output it.

[0051] 7, the product planning department 120 can predict the expected sales volume by date (time) based on input data for a specific product and output it by date (over time). For example, the cumulative sales volume (a) by time for the input product can be displayed in a graph.

[0052] According to one embodiment, the product planning unit 120 may search for one or more similar products based on the predicted input data. For example, the product planning unit 120 may search for one or more products having similar conditions to the product to be predicted based on at least one of a product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price. For example, the product planning unit 120 may calculate a similarity based on the product design image, product information, distribution platform, and sales price, and search for a predetermined number of highly similar products as similar products.

[0053] According to one embodiment, the product planning department 120 extracts comparison input data corresponding to the prediction input data for one or more similar products and estimates a predicted sales volume based on the comparison input data for each of the one or more similar products. The prediction input data is input data for estimating a predicted sales volume for a product desired by a user, while the comparison input data is input data related to products to be compared, including products similar to the product desired by the user.

[0054] For example, the product planning department 120 may search for a predetermined number of similar products that are highly similar, and may search for information regarding at least one of the product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price for the relevant product. Through this, the product planning department 120 may estimate the expected sales volume of the similar products based on the information regarding the searched similar products. For example, as shown in FIG. 7, the product planning department 120 may search for two similar products (b, c) and predict the expected sales volume for each of the two similar products (b, c). In this case, the product planning department 120 may predict the expected sales volume for the two similar products (b, c) using the above-described demand forecasting model.

[0055] According to one embodiment, the product planning unit 120 may search for actual sales data for one or more similar products and output the data along with estimated sales volumes for each of the one or more similar products. For example, in the case of two similar products (b, c), where data on past sales exists, the product planning unit 120 may search for actual sales data information for the corresponding products and output the searched actual sales volumes. For example, as shown in FIG. 7, a graph (b', c') of actual sales volumes may be output.

[0056] According to one example, the product planning department 120 can predict additional sales volume based on actual sales volume data. For example, as shown in FIG. 7, graph (c') shows that actual sales may be interrupted at a certain point (star). This may be due to a product being sold out or other reasons, and the product may not be able to be sold despite demand. In such a case, the product planning department 120 can predict the quantity that can be sold if the product is sold for an additional period after the interruption based on the actual sales volume data, and display this on the graph (c').

[0057] Referring to FIG. 7, a difference may occur between the predicted product sales volume (b, c) and the actual sales volume (b', c'). This may be due to a difference between the input data and the data generated during actual sales, or due to uninputted data acting as a variable. For example, in the case of product b, marketing may be carried out according to the planned marketing schedule, but an unexpected incident may occur, such as a celebrity posting about the product on their social media account. In such a case, a phenomenon that differs from the input data may occur, resulting in a sales volume that differs from the predicted sales volume.

[0058] According to one example, the product planning unit 120 may output comparison input data corresponding to predicted input data for one or more similar products. For example, the product planning unit 120 may output, as the comparison input data, information regarding at least one of a product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price for the similar products.

[0059] For example, the product planning unit 120 may search for and output information on data that may change over time, unlike initially input data. For example, a product's image, sales price, etc. may be fixed data, but celebrity scores, social media scores, etc. may be variables generated by external consumers and may change over time.

[0060] As an example, the product planning department 120 can search for data by time period for similar products, calculate celebrity scores, social media scores, etc., and output them. For example, as shown in Figure 8, the product planning department 120 can output the celebrity scores and social media scores by time period for product b along with the predicted sales quantity (b) and actual sales quantity (b'). As shown in Figure 8, it can be seen that the celebrity scores and social media scores appear high at specific times, and the user can make predictions about variables by confirming that the sales quantity increased sharply at those times.

[0061] FIG. 9 is a flowchart showing a product planning support method according to an embodiment.

[0062] According to one example, the product planning support device may be a computing device having one or more processors and a memory that stores one or more programs executed by the one or more processors.

[0063] According to one embodiment, the product planning support device may classify one or more search images retrieved based on at least one of the one or more collected keywords into one or more groups (910), and may receive a reference image selection input from a user to select at least one of the one or more search images classified into the one or more groups (920). The product planning support device may also receive image editing prompt data from the user (930), and may output a product design image based on the one or more images selected in the reference image selection input and the image editing prompt data (940).

[0064] Among the embodiments in FIG. 9, those that overlap with the contents explained with reference to FIGS. 1 to 8 are omitted.

[0065] One aspect of the present invention can be embodied as computer-readable code on a computer-readable recording medium. Codes and code segments embodying the program can be easily construed by a computer programmer skilled in the art. The computer-readable recording medium can include all types of recording devices in which data readable by a computer system is stored. Examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical disk, etc. Furthermore, the computer-readable recording medium can be distributed across computer systems connected via a network, and the computer-readable code can be created and executed in a distributed manner.

[0066] The present invention has been described above with reference to preferred embodiments. Those skilled in the art will understand that the present invention can be embodied in various modified forms without departing from the essential characteristics of the present invention. Therefore, the scope of the present invention is not limited to the above-described embodiments, but should be interpreted to include various embodiments within the scope equivalent to the claims. [Explanation of symbols]

[0067] 100: Product planning support device 110: Interface section 120: Product Planning Department

Claims

1. an interface unit for data input and output; and a product planning unit connected to the interface unit, The product planning department Classifying one or more search images searched based on at least one of the collected one or more keywords into one or more groups; receiving a reference image selection input from a user through the interface unit to select at least one of the one or more search images classified into the one or more groups; receiving image editing prompt data from a user through the interface unit; a product planning support device that outputs a product design image based on the one or more images selected by the reference image selection input and the image editing prompt data;

2. The product planning department Classifying the collected one or more keywords into broad classification keywords and sub-classification keywords according to a predetermined rule; The product planning support device according to claim 1 , further comprising: receiving a user's keyword selection input for at least one of a major classification keyword and a sub-classification keyword through the interface unit; and outputting search volume by date based on the user's keyword selection input.

3. The product planning department receiving an input for at least one of a search platform, a search keyword, and a search period through the interface unit; Searching for images from the search platform based on the search keyword and the search period to generate one or more search images; The product planning support device of claim 1 , further comprising: categorizing the one or more search images to generate one or more crowds.

4. The product planning department receiving an input from a user for a crowd size; determining a predetermined classification criterion based on the number of clusters; The product planning support device according to claim 3 , wherein the number of crowds is generated based on a predetermined classification criterion determined by the number of crowds.

5. The product planning department 2. The product planning support device of claim 1, wherein the expected sales volume is estimated based on prediction input data including at least one of the product design image, product information, distribution platform, celebrity score, social media score, sales start date, and sales price.

6. 6. The product planning support device according to claim 5, wherein the expected sales volume is sales volume data by date.

7. The product planning department The product planning support device according to claim 5 , wherein one or more similar products are searched for based on the predicted input data.

8. The product planning department extracting comparison input data corresponding to the prediction input data for the one or more similar products; 8. The product planning support device according to claim 7, wherein the expected sales volume is estimated based on comparison input data for each of the one or more similar products.

9. The product planning department 9. The product planning support device according to claim 8, further comprising: retrieving actual sales data for the one or more similar products and outputting the data together with an estimated forecast sales volume for each of the one or more similar products.

10. The product planning department 8. The product planning support device according to claim 7, wherein comparison input data corresponding to the predicted input data for the one or more similar products is output.

11. one or more processors, and 1. A method performed on a computing device having a memory storing one or more programs to be executed by the one or more processors, comprising: classifying the one or more retrieved images into one or more groups based on at least one of the collected one or more keywords; receiving a reference image selection input from a user selecting at least one of the one or more search images categorized into the one or more groups; receiving image editing prompt data from a user; and The product planning support method includes a step of outputting a product design image based on one or more images selected by the reference image selection input and the image editing prompt data.

Images