TWM591213U - Auxiliary device for merchandising based on image recognition - Google Patents

Abstract

A commodity sales assistant device based on image recognition includes a camera, a database and a processing module. The camera is used to obtain the target image. The database is used to store multiple financial commodity information and multiple accessory image information. The processing module is electrically connected to the camera and the database. The processing module is used to obtain a set of biometrics from the target image and generate potential financial demand information based on the set of biometrics. The processing module is further used to extract accessories from the target image Image, and the value of the accessory is estimated based on one or more accessory features of the accessory image and the image information of the accessory, based on which the estimated income interval information is generated, and the processing module is based on the potential financial demand information and the estimated The estimated income interval information searches and outputs one or more financial commodity information among the financial commodity information in the database.

Description

Auxiliary device for merchandising based on image recognition

The present invention relates to an auxiliary device for merchandising based on image recognition, in particular to an auxiliary device for merchandising using image recognition of biological and accessories.

As the competition in the financial industry market becomes fiercer, banks always hope that their employees can vigorously promote financial products, such as credit cards, cash cards, deposits, futures, funds, etc., with a view to expanding their own businesses and improving their operating conditions.

However, in general, when a bank teller serves an unfamiliar customer, it often can only rely on personal experience to initially determine the unfamiliar customer's earning capacity and demand, and cannot accurately grasp the information through multiple image analysis within a short counter service time. The property of a stranger customer, which in turn leads to the loss of the opportunity to sell their financial products. Therefore, in this field, there is a need for a highly accurate multivariate image analysis device that can assist tellers to quickly understand customer attributes and recommend appropriate financial products accordingly.

The present invention proposes a commodity marketing auxiliary device based on image recognition, which comprehensively judges the character's attributes through the character's biometric image recognition and the accessory image recognition, and then promptly recommends suitable financial products.

According to an embodiment of the present invention, an auxiliary device for merchandising based on image recognition is disclosed, which includes a camera, a database and a processing module. The camera is used to obtain the target image. The database is used to store multiple financial commodity information and multiple accessory image information. The processing module is connected to the camera and the database. The processing module is used to obtain a set of biometrics from the target image and generate potential financial demand information based on the set of biometrics. The processing module is further used to capture the image of the accessory from the target image , And estimate the value of the accessory based on the feature or features of the accessory’s image and the image information of the accessory, to generate estimated revenue interval information based on it, and the processing module based on the potential financial demand information and estimated revenue The interval information searches and outputs one or more financial commodity information in the financial commodity information in the database.

In summary, in the merchandising assistant device based on image recognition proposed by the new model, the biometrics are obtained through facial image recognition of the person, and the potential financial needs of the customer are analyzed according to the biometrics. Identify and obtain the characteristics of accessories, and evaluate the value of accessories worn by customers based on the characteristics of accessories to determine the customer's income range, and then combine the above two conditions to determine the attributes of customers, so that the teller can promptly recommend suitable for the limited counter service time The customer's financial products can effectively increase the probability of successful financial product recommendation.

The above description of the content of the disclosure and the following description of the embodiments are used to demonstrate and explain the spirit and principle of the new model, and provide a further explanation of the patent application scope of the new model.

The following describes in detail the detailed features and advantages of the new model in the embodiments. Its content is sufficient for any person skilled in the relevant art to understand and implement the technical content of the new model, and according to the contents disclosed in this specification, the scope of patent application and the drawings Anyone who is familiar with related skills can easily understand the purpose and advantages of this new type. The following examples further illustrate the point of view of the novel, but do not limit the scope of the novel with any point of view.

Please refer to FIG. 1. FIG. 1 is a functional block diagram of an image recognition-based product sales assistant device according to an embodiment of the present invention. As shown in FIG. 1, the product sales support device 1 includes a camera 10, a database 11 and a processing module 12. The camera 10 has a camera host and is externally connected with one or more camera lenses (not shown) for capturing a person to obtain a target image, and the target image is returned to the host in the camera 10, where the target image is, for example Images of bank customers. The database 11 is used to store multiple financial commodity information. In practice, the database 11 is built in the memory or storage unit of the device. The database 11 can store various types of financial commodity information, such as credit cards, cash cards, deposits, futures, funds, stocks, bonds, etc. .

The processing module 12 is connected to the camera 10 and the database 11. The processing module 12 is used to obtain a set of biometrics from the target image (such as a bank customer) captured by the camera 10, and the biometrics include the client's hair status and skin At least one of status, facial features, posture, or body language characteristics. In a practical example, after the processing module 12 receives the target image from the camera 10, the image processor 121 inside the processing module 12 will first capture a face image in the target image, and then remove the target image from the target image. The area belonging to the face image is regarded as the background and filtered out. Next, the image processor 121 of the processing module 12 obtains a plurality of feature points in the face image to obtain a sub-image composed of the feature points.

More specifically, the area enclosed by the connection of these feature points in the face image is the sub-image, where the sub-image includes one of the facial features of the person and the surrounding skin The image is information about the person’s biological characteristics. In this embodiment, the image processor 121 of the processing module 12 can obtain more sub-images by obtaining more feature points in the face image, for example, to obtain all facial features of the person and the corresponding peripheral skin The image serves as information about the person’s biological characteristics.

When the computing unit 122 of the processing module 12 obtains images of all facial features from the image processor 121 and their corresponding peripheral skin, the person can be estimated through the appearance of the facial features and the surrounding skin state (ie, biological characteristics) Age and/or gender. In detail, the facial features include, for example, the shape of the eyes, the distance between the eyes, the length and thickness of the eyebrows, the width of the nose, the length of the bridge of the nose, the thickness of the lips, or the shape of the ears. Quantity or color and quantity of skin markings, etc.

The arithmetic unit 122 of the processing module 12 further generates potential financial demand information according to the age and/or gender corresponding to the set of biological characteristics. In practice, the arithmetic unit 122 may include, for example, a processor, a microprocessor, a controller, a microcontroller, and other components with arithmetic functions. Information on various types of financial commodities in the database 11 can be pre-classified into high-risk short-term investment commodities (such as stocks or futures), low-risk long-term investment commodities (such as time deposits), and general commodities (such as credit or cash cards), for example . Further, the computing unit 122 of the processing module 12 can collect and statistically analyze customer information within the bank, and accordingly set the financial products that customers of different ages and ages tend to require, as a basis for generating information on potential financial needs.

In more detail, the arithmetic unit 122 of the processing module 12 can analyze the customer's gender and/or age based on biological characteristics to determine its potential financial demand information. In fact, customers of different genders and ages may have demand for financial products There are differences, for example, the investment concept of men is more positive than that of women. People in the elderly age group are more conservative in their financial management than those in the middle age group, while people in the younger age group may not have the idea of investment finance at all. In an embodiment, the database 11 stores a list of financial needs, corresponding to different types of demand information under different gender and/or age conditions. The different types of demand information may include first-type demand information and second-type demand information The third type of demand information corresponds to the aforementioned high-risk short-term investment commodities, low-risk long-term investment commodities and general commodities, respectively. The arithmetic unit 122 of the processing module 12 can analyze the gender and/or age of the customer according to the biological characteristics, and select the first type of demand information, the second type of demand information, and the third type in the financial demand list according to the gender and/or age One of the demand information is used as potential financial demand information.

Using the above concept as a benchmark, and using practical examples to illustrate, if the customer is a female and is about 55 years old (such as a housewife), the investment concept is more conservative, and it is more acceptable to accept low-risk financial products that can benefit from long-term investment. Therefore, the arithmetic unit 122 of the processing module 12 can determine that the potential financial demand information of the customer is the second type of demand information. If the client is a male and is about 38 years old (for example, a small and medium-sized enterprise executive), it is evaluated that its investment concept is more open, and it is more acceptable to accept high-risk financial products that can benefit from short-term investment. Therefore, the arithmetic unit 122 of the processing module 12 can be Judge that the customer's potential financial demand information is the first type of demand information. If the customer is a male (or female) and is about 23 years old (for example, a freshman in the workplace), it is estimated that his interest in investment and financial management is not high or no idea, and it is more suitable for general financial products (credit card or cash card), so the processing module The arithmetic unit 122 of 12 can determine that the customer's potential financial demand information is the third type of demand information. The processor 122 of the processing module 12 can estimate the types of financial commodities that the customer may need by analyzing the gender and age of the customer.

However, according to gender and age, it is only possible to estimate which category of demand the customer's potential demand for financial commodities belongs to. It is not enough to determine whether the customer's financial ability is consistent with the investment unit price of this category of financial commodities. For example, suppose that the customer is suitable for the second type of demand information (suitable for investing in high-risk products) based on the customer's gender and age, such as stocks. However, the unit price of stocks of different listed companies is not the same, and the unit price of stocks of some listed companies may reach millions. If the income ability of the customer is unaffordable, it is obviously inappropriate to recommend such stocks to the customer. Therefore, it is necessary to further evaluate the amount of income of customers before they can select and output the truly appropriate financial commodity information from multiple financial commodity information in the corresponding category and recommend it to the customer.

In view of this, the processing module 12 further captures the image of the accessory from the target image, and the processing module 12 further pre-processes the image of the accessory according to one or more accessory features and the image information of the accessory in the database 11 Estimate the value of the accessory and use it to generate estimated revenue range information to assess the customer's income. Then, the processing module 12 searches and outputs one or more financial commodity information in the financial commodity information in the database 11 according to the potential financial demand information and the estimated income interval information. In practice, the accessories are, for example, watches, wallets, neckties/scarves, necklaces, earrings, etc. worn or carried on customers. In one embodiment, the feature of the accessory includes an outer edge outline, a brand trademark, or a texture pattern. More specifically, the outline of the outer edge is, for example, a necklace or jewelry of a particular shape, the brand trademark is, for example, a brand logo on a watch or a wallet, and the pattern is, for example, the pattern on a tie or scarf. .

In one embodiment, the image processor 121 of the processing module 12 will first search and capture an accessory image in the target image, and filter out the region of the accessory image that does not belong to the accessory image as a background and filter it out . Next, the image processor 121 of the processing module 12 obtains a plurality of feature points in the accessory image to obtain a sub-image composed of the feature points. Specifically, the area enclosed by the connection of the feature points in the accessory image is the sub-image, for example, a rectangular area or an area of another shape is enclosed as the sub-image according to the connection of the feature points. Wherein, the sub-image will include the image of the product logo of the accessory and the special patterns around it, that is, the information of the feature of the accessory.

Taking the watch as an example for illustration, the image processor 121 of the processing module 12 obtains a plurality of characteristic points of the outline of the watch body in the image of the watch, and uses these characteristic points to enclose a rectangular area or an area of another shape as a sub-image . In this case, the sub-images include product logos and special patterns embedded on the surface of the watch body. Therefore, the image processor 121 of the processing module 12 can extract the product logo and the special pattern as accessory features.

In this embodiment, the database 11 stores various types of accessory image information, which includes trademark patterns, pattern designs or outer contours of major brand products on the market. The accessory image information is based on their corresponding brands in The generally accepted value/price on the market is pre-classified as the first-level accessory image, the middle-level accessory image and the high-level accessory image from low to high, as the basis for judging the value of the accessory. In other words, the processor 122 of the processing module 12 compares the accessory features from the image processor 121 one by one to the primary accessory image, the intermediate accessory image and the high-end accessory image to determine whether the accessory feature is consistent Which one of the first-, middle-, or high-end accessory images can be used to determine whether the value of the accessory is the first-, middle-, or high-end level. Then, the arithmetic unit 122 of the processing module 12 selects one of a plurality of income intervals as the estimated income interval information according to the judgment result. In practice, for example, the arithmetic unit 122 of the processing module 12 can be set to make the judgment criterion be that the initial level corresponds to the annual income range of about 600,000 yuan or less, and the intermediate level corresponds to the annual income range of about 600,000 to one Between one million yuan, and the high-grade level corresponds to an annual income range of about one million yuan to several million yuan.

Please further refer to FIG. 2, which is a schematic diagram of a target image according to an embodiment of the present invention. The following uses the embodiment of FIG. 2 in conjunction with FIG. 1 as a practical example. As shown in FIG. 2, it is assumed that the target image A0 acquired by the camera 10 through shooting, and the processing module 12 with the image processor 121 and the arithmetic unit 122 through biological features Extraction and analysis determine that the client's gender is male and the age is about 35 to 40 years old. Further evaluation of its potential financial needs may be the first type of demand information, such as high-risk financial products such as stocks or futures. In addition, since the customer wears a watch of a certain brand as an accessory, the processing module 12 further recognizes the watch based on the comparison result of the brand trademark on the watch and the pre-stored accessory image information in the image A1 of the accessory The brand of the watch corresponds to an image of a high-end accessory in the database 11, so it is judged that the value of the watch is quite expensive (it is a high-end accessory). Based on this, the processing module 12 can infer that the customer's estimated revenue range is quite high, which is approximately between one million and several million yuan in annual revenue, so the corresponding annual revenue is one million to several million yuan. Estimated income range information. Based on the above two estimated conditions (ie, potential financial demand and estimated income range information), the processing module 12 can search the database 11 for high-risk financial products such as stocks or futures with relatively high unit prices to recommend to the customer.

In practice, considering that the estimated revenue range information of a customer’s single accessory may not be accurate enough, so in one embodiment of the present invention, the processing module 12 of the merchandising assistant 1 may be based on the same The method further captures an image of another accessory from the target image, and the processing module 12 estimates the value of the other accessory according to one or more accessory features of the image of the other accessory, further The arithmetic unit 122 of the processing module 12 adjusts the estimated income interval information according to the value of another accessory and the value of the accessory, and then the arithmetic unit 122 of the processing module 12 can adjust the estimated income interval information according to the potential financial demand information , Search and output suitable financial commodity information in database 11. In practice, the processing module 12 may adjust the estimated income interval information to increase or decrease the estimated income interval. In the embodiment of FIG. 2, another accessory of the customer is a wallet, and the processing module 12 compares the accessory image in the database 11 according to one or more accessory features of the wallet accessory image A2 Information, for example, the pattern on the wallet matches an image of an initial accessory, the processing module 12 can estimate that the wallet is only a cheap commodity (belonging to an initial level accessory), and its value is not particularly expensive. Therefore, the processing module 12 makes a comprehensive judgment based on the value of the primary low-end wallet and the value of the aforementioned high-end high-end watches, and adjusts the original estimated income range to approximately 600,000 to 1 million yuan per year. That is to say, in this embodiment, the processing module 12 determines that the customer’s annual revenue is not as high as originally estimated according to the value of another accessory. Therefore, after the estimated income range is lowered, the processing module 12 will The database 11 searches for high-risk financial commodities such as stocks or futures with an intermediate unit price and recommends to the client.

Please refer to FIG. 1. The merchandising assistant device 1 shown in FIG. 1 further includes a caller 13, which is electrically connected to the processing module 12. In this embodiment, the numbering machine 13 is electrically connected to each counter-end device 21 to 23 for each bank teller to notify the next customer to go to the designated counter to handle the business. The caller 13 is used to generate a service serial number corresponding to a counter-end device when triggered, and the arithmetic unit 122 of the processing module 12 can transmit one or more financial commodity information to the corresponding counter-end device according to the service serial number. In the embodiment of FIG. 1, each counter-side device 21, 22, and 23 is respectively equipped with a camera lens and connected to the camera, and each counter-side device 21, 22, and 23 is assigned a service number 001, 002, and 003. In an implementation scenario, assuming that the teller at the counter-end device 22 wants to serve the next customer, he can operate the counter-end device 22 to send a trigger signal to the caller 13, at which time the caller 13 executes the call and generates a corresponding Service serial number 002 and transmits the service serial number 002 to the arithmetic unit 122 of the processing module 12.

After the caller 13 executes the call, the customer will move to the corresponding counter device 22, and the camera 10 obtains the customer's image as the target image through the corresponding camera lens, and the calculation of the processing module 12 Based on the target image, the device 122 judges the customer's potential financial needs and estimated income range through the aforementioned biological analysis and accessory value assessment, and combines the estimated conditions of the two to search one or more suitable financial commodity information in the database for promotion . Then, the arithmetic unit 122 of the processing module 12 can send the information of one or more financial commodities suggested to the corresponding counter device 22 according to the service serial number 002. Therefore, the teller can promptly sell the financial commodity information displayed on the screen of the counter-end device 22 to the customer, so that the probability of successful sales can be greatly improved.

Please further refer to FIG. 3, which is a flowchart of a merchandise sales assistance method for a merchandise sales assistance device based on image recognition according to an embodiment of the present invention. This method is applicable to the financial product sales assistance device of FIG. 1. As shown in FIG. 3, in step S1, the camera 10 acquires the target video. In step S2, the processing module 12 obtains a set of biometrics from the target image, and generates potential financial demand information according to the set of biometrics. In step S3, the processing module 12 captures the image of the accessory from the target image, and estimates the value of the accessory based on the one or more accessory features and the plurality of accessory image information possessed by the image of the accessory, thereby generating a prediction Estimated income range information. In step S4, the processing module 12 searches and outputs one or more financial commodity information among the multiple financial commodity information in the database according to the potential financial demand information and the estimated income interval information.

Please refer to FIG. 4. FIG. 4 is a flow chart of another method for assisting merchandise promotion based on an image recognition merchandise assisting device according to an embodiment of the present invention. The steps S11-S13 and S16 of the embodiment of FIG. 4 are similar to the steps S1-S3 and S4 of the embodiment of FIG. 3, but the difference is that the embodiment of FIG. 4 further includes steps S14 and S15. In step S14, the processing module 12 captures an image of another accessory from the target image, and estimates the other accessory according to the feature or features of the accessory and the image information of the accessories The value of an accessory. In step S15, the processing module 12 adjusts the estimated income interval information according to the value of the accessory and the value of the other accessory.

Please refer to FIG. 5. FIG. 5 is a flow chart of another method for assisting merchandise promotion based on an image recognition merchandise assisting device according to an embodiment of the present invention. Steps S111 to S114 in the embodiment of FIG. 5 are similar to steps S1 to S4 in the embodiment of FIG. 3. The difference is that FIG. 5 further includes that before the camera 10 acquires the target image, in step S110, when the caller 13 is triggered, the caller 13 generates and transmits the service serial number to the processing module 12. After the processing module 12 outputs the one or more financial commodity information, in step S115, the processing module 12 transmits the output one or more financial commodity information to the counter device corresponding to the service serial number according to the service serial number .

In one embodiment, the processing module 12 estimates the value of the accessory based on the one or more accessory features and the image information of the accessory, and generating the estimated revenue interval information based on the processing module Group 12 obtains the outer contour, brand trademark or texture pattern of the image of the accessory as the one or more accessory features and compares each accessory feature to the accessory image information individually to select one of the multiple income intervals as Estimated income range information.

In summary, in the merchandising assistant device based on image recognition proposed by the new model, the biometrics are obtained through facial image recognition of the person, and the potential financial needs of the customer are analyzed according to the biometrics. Identify and obtain the characteristics of accessories, and evaluate the value of the accessories worn by the customer based on the characteristics of the accessories to determine the customer's income range, and then combine the above two conditions to determine the customer's attributes to help the teller recommend the suitable customer within the limited counter service time Financial commodities, this can effectively increase the probability of successful financial commodities recommendation.

Although the present invention is disclosed as above with the foregoing embodiments, it is not intended to limit the present invention. Without departing from the spirit and scope of the new model, all modifications and retouchings belong to the patent protection scope of the new model. For the protection scope defined by this new model, please refer to the attached patent application scope.

1‧‧‧Commodity sales assistant 10‧‧‧Camera 11‧‧‧ Database 12‧‧‧Processing module 121‧‧‧Image processor 122‧‧‧Calculator 13‧‧‧Caller 21~23‧‧‧ Counter device A0‧‧‧Target image Images of A1, A2‧‧‧Accessories

FIG. 1 is a functional block diagram of an image recognition-based product sales assistant device according to an embodiment of the present invention. FIG. 2 is a schematic diagram of a target image according to an embodiment of the present invention. FIG. 3 is a flowchart of a merchandise sales assistance method to which a merchandise sales assistance device based on image recognition is applied according to an embodiment of the present invention. FIG. 4 is a flow chart of another method for assisting merchandise promotion based on an image recognition merchandise assisting device according to an embodiment of the present invention. FIG. 5 is a flowchart of another method for assisting product sales based on an image recognition-based product sales support device according to an embodiment of the present invention.

1‧‧‧Commodity sales assistant

10‧‧‧Camera

11‧‧‧ Database

12‧‧‧Processing module

121‧‧‧Image processor

122‧‧‧Calculator

13‧‧‧Caller

21~23‧‧‧ Counter device

Claims (4)

A commodity marketing auxiliary device based on image recognition, including: a camera for obtaining a target image; A database for storing multiple financial commodity information and multiple accessory image information; and A processing module is connected to the camera and the database. The processing module is used to obtain a set of biometrics from the target image and generate a potential financial demand information according to the set of biometrics. The processing module is further used to Capture an image of an accessory from the target image, and estimate the value of the accessory based on one or more accessory features of the accessory image and the image information of the accessories, thereby generating an estimated income interval information, The processing module searches and outputs one or more financial commodity information in the financial commodity information in the database according to the potential financial demand information and the estimated income interval information. The merchandising assistant device based on image recognition according to claim 1, wherein the processing module further captures an image of another accessory from the target image and has one or more accessories according to the image of the other accessory The features and the image information of the accessories estimate the value of the other accessory, and the processing module adjusts the estimated income interval information according to the value of the other accessory and the value of the accessory. The merchandising assistant device based on image recognition as described in claim 1, wherein the one or more accessory features include an outer edge outline, a brand trademark or a pattern design. The auxiliary device for merchandise promotion based on image recognition as described in claim 1 further includes a caller electromechanically connected to the processing module, and the caller is used to generate a service serial number corresponding to a counter device when triggered. The processing module transmits the one or more financial commodity information to the corresponding counter device according to the service serial number.

Images