Method for automatically identifying television picture articles and associating corresponding commodities for selling

Abstract

The invention provides a method for automatically identifying television picture articles and associating with corresponding commodities for selling, and relates to the technical field of computer vision identification. The method for automatically identifying the television picture articles and associating the corresponding commodities for selling comprises the following steps of S1, training a commodity object identification model by utilizing an existing commodity database; and S2, vectorizing all commodities by using the identification model, and storing the vectorized commodities in a vector database. In the invention, through the object detection model, video stream processing, vector recall and other modes, the self-association of the video and the commodity is realized, human participation is not needed, the automation capability is relatively high, when the user watches the interested video, the commodity concerned by the user appears, the purchase desire of the user can be improved, the purchase process is reduced, and the user experience is improved. And platform operation personnel do not need to configure related commodities for the television column, and all processes are carried out by the system, so that the system is more convenient to use.

Description

Technical Field

[0001] This invention relates to the field of computer vision recognition technology, specifically a method for automatically identifying items on a television screen and associating them with corresponding products for sale. Background Technology

[0002] With the development of technology, smart TVs have become widely popular. Smart TVs not only play traditional television programs but also possess powerful computing and network connectivity capabilities. They can install various applications, providing the hardware foundation for functions such as automatically recognizing items and selling goods on the TV.

[0003] To achieve the function of automatically identifying items and associating them with product sales, a well-designed interface and interaction method are needed in the TV software. The software interface includes interfaces with the image recognition module, interfaces with the product database, and interfaces for user interaction. For example, the image recognition module needs to pass the recognition results to the product association module, which requires defining the data transmission format and interface.

[0004] Currently, in most television programs, the process of selling goods is handled manually by staff who configure the program and the goods, which is cumbersome and inefficient.

[0005] Therefore, the present invention aims to provide a method that can automatically identify items on television screens and associate them with corresponding products for sale, thereby automatically and quickly identifying products appearing in television programs and associating them with existing products for sale, in order to solve the problems mentioned in the background art. Summary of the Invention

[0006] (a) Technical problems to be solved

[0007] To address the shortcomings of existing technologies, this invention provides a method for automatically identifying items on television screens and associating them with corresponding products for sale. This solves the problem that in most television programs currently on the market, the process of selling products is cumbersome and inefficient, as the program and product settings are manually configured.

[0008] (II) Technical Solution

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A method for automatically identifying items on a television screen and associating them with corresponding products for sale includes the following steps:

[0011] Step S1. Train a product object recognition model using the existing product database;

[0012] Step S2. Vectorize all products using the recognition model and store them in the vector database;

[0013] Step S3. Integrate and embed the object recognition model into the television set. When the television is playing video streams, use the model to recognize objects in the television screen.

[0014] Step S4. Recognize the captured object image, convert it into a vector, and send it to the terminal server;

[0015] Step S5. The terminal server completes vector matching and finally identifies the corresponding product;

[0016] Step S6. The terminal server pushes the corresponding product information to the TV for display, thus completing the automatic association of products.

[0017] Furthermore, the product database in step S1 contains detailed product information, such as product name, brand, model, price, inventory, and purchase link. The database structure can be designed according to actual needs. For example, a relational database can be used to store product category information, attribute information, and sales information in different tables and link them through primary keys and foreign keys.

[0018] Furthermore, in step S1, the construction of the object detection and recognition model requires collecting a large number of images containing various commodities and labeling the images to indicate the category of each item. This can be done manually or by using some existing labeling tools. The model can learn the features of the images and can identify key features such as the shape, texture, and color of different items.

[0019] Furthermore, step S3, which utilizes this model to identify objects in the television screen, specifically includes:

[0020] 1) Obtaining picture frames from a television signal source can be achieved through a software or hardware interface with the television;

[0021] 2) After acquiring the image, preprocess the image, including adjusting the image size, color correction, and noise reduction.

[0022] 3) Input the preprocessed image frames into the trained object recognition model, and the model will output the categories of objects identified in the image and their probabilities;

[0023] 4) Set an appropriate probability threshold, usually 60%-70%. Only items with a recognition probability higher than this threshold will be further processed.

[0024] Furthermore, in step S5, the terminal server performs a query and matching in the product database based on the identified item category. It can use the item category name as a keyword to perform a fuzzy or precise query in the product name or category field of the database, and finally identify the corresponding product.

[0025] Furthermore, in step S6, the associated product information is displayed to the user in an intuitive way, which can be done by displaying the product's image, name, price, and purchase link information on one side of the TV screen or through a pop-up window.

[0026] Furthermore, after step S6, the television also provides user interaction functions, such as allowing users to view more product details, add products to the shopping cart, and select the quantity to purchase via remote control or voice control.

[0027] Furthermore, the television can guide users through the purchase process when they decide to buy a product. If the purchase is redirected to an external e-commerce platform, it needs to interface with the e-commerce platform to ensure secure redirection and transmission of user information. For some simple products, the payment process can also be completed directly on the television.

[0028] (III) Beneficial Effects

[0029] This invention provides a method for automatically identifying items on a television screen and associating them with corresponding products for sale. It has the following beneficial effects:

[0030] 1. This invention provides a method for automatically identifying items on television screens and associating them with corresponding products for sale. Through object detection models, video stream processing, and vector recall, it achieves self-association between videos and products without human intervention, possessing a high degree of automation. It can display products that users are interested in when watching videos of interest, thereby increasing users' desire to buy, reducing the purchase process, and eliminating the need for platform operators to configure related products for television programs, as everything is handled by the system, making it more convenient to use.

[0031] 2. This invention provides a method for automatically identifying items in a television screen and associating them with corresponding products for sale. It employs a high-precision object recognition algorithm model, which can efficiently and accurately process frame images appearing in a video stream. Furthermore, based on a high-performance vector database, it can store massive amounts of image vector data, supporting efficient vector retrieval and similarity calculation. Based on user behavior, preferences, and other actions, it intelligently recommends products of interest to users. Attached Figure Description

[0032] Figure 1 This is a flowchart of a method for automatically identifying items on a television screen and associating them with corresponding products for sale, according to the present invention. Detailed Implementation

[0033] The technical solutions of the specific embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described specific embodiments are only a part of the specific embodiments of the present invention, and not all of them. Based on the specific embodiments of the present invention, all other specific embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] like Figure 1 As shown in the figure, a specific embodiment of the present invention provides a method for automatically identifying items on a television screen and associating them with corresponding goods for sale, including the following steps:

[0035] Step S1. Train a product object recognition model using the existing product database;

[0036] The product database contains detailed product information, such as product name, brand, model, price, inventory, and purchase link. The database structure can be designed according to actual needs. For example, a relational database can be used to store product category information, attribute information, and sales information in different tables and link them through primary keys and foreign keys.

[0037] Building an object detection and recognition model requires collecting a large number of images containing various products and labeling the images to indicate the category of each item. This can be done manually or by using some existing labeling tools. The model can learn the features of the images and recognize key features such as the shape, texture, and color of different items.

[0038] Step S2. Vectorize all products using the recognition model and store them in the vector database;

[0039] Step S3. Integrate and embed the object recognition model into the television set. When the television is playing a video stream, the model is used to recognize objects in the television screen. Specifically, this includes:

[0040] 1) Obtaining picture frames from a television signal source can be achieved through a software or hardware interface with the television;

[0041] 2) After acquiring the image, preprocess the image, including adjusting the image size, color correction, and noise reduction.

[0042] 3) Input the preprocessed image frames into the trained object recognition model, and the model will output the categories of objects identified in the image and their probabilities;

[0043] 4) Set an appropriate probability threshold, usually 60%-70%. Only items with a recognition probability higher than this threshold will be further processed.

[0044] Step S4. Recognize the captured object image, convert it into a vector, and send it to the terminal server;

[0045] Step S5. The terminal server completes vector matching and finally identifies the corresponding product;

[0046] The terminal server performs a query and matching in the product database based on the identified item category. It can use the item category name as a keyword to perform a fuzzy or precise query in the product name or category field of the database, and finally identify the corresponding product.

[0047] Step S6. The terminal server pushes the corresponding product information to the TV for display, thus completing the automatic association of products;

[0048] The associated product information can be displayed to users in an intuitive way, such as on the side of the TV screen or through a pop-up window, showing the product's image, name, price, and purchase link.

[0049] The television also offers user interaction features, such as allowing users to view more product details, add products to their shopping cart, and select the quantity to purchase via remote control or voice control.

[0050] Televisions can also guide users through the purchase process when they decide to buy a product. If the purchase is redirected to an external e-commerce platform, it needs to be integrated with the e-commerce platform's interface to ensure secure redirection and transmission of user information. For some simple products, the payment process can also be completed directly on the television.

[0051] In this invention, through object detection models, video stream processing, and vector recall, the self-association of videos and products is achieved without human intervention, exhibiting a high degree of automation. When users watch videos of interest, products that users care about will appear, which can increase users' desire to buy, reduce the purchase process, and eliminate the need for platform operators to configure related products for TV programs, as everything is handled by the system, making it more convenient to use.

[0052] This invention employs a high-precision object recognition algorithm model, which can efficiently and accurately process frame images appearing in video streams. Furthermore, based on a high-performance vector database, it can store massive amounts of image vector data, supporting efficient vector retrieval and similarity calculation. Based on user behavior, preferences, and other actions, it intelligently recommends products of interest to users.

[0053] Although specific embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these specific embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for automatically identifying items on a television screen and associating them with corresponding products for sale, characterized in that, Includes the following steps: Step S1. Train a product object recognition model using the existing product database; Step S2. Vectorize all products using the recognition model and store them in a vector database; Step S3. Integrate and embed the object recognition model into the television set. When the television is playing video streams, use the model to recognize objects in the television screen. Step S4. Recognize the captured object image, convert it into a vector, and send it to the terminal server; Step S5. The terminal server completes vector matching and finally identifies the corresponding product; Step S6. The terminal server pushes the corresponding product information to the TV for display, thus completing the automatic association of products.

2. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... The product database in step S1 contains detailed product information, such as product name, brand, model, price, inventory, and purchase link. The database structure can be designed according to actual needs. For example, a relational database can be used to store product category information, attribute information, and sales information in different tables and link them through primary keys and foreign keys.

3. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... In step S1, the construction of the object detection and recognition model requires collecting a large number of images containing various commodities and labeling the images to indicate the category of each item. This can be done manually or by using some existing labeling tools. The model can learn the features of the images and identify key features such as the shape, texture, and color of different items.

4. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... The specific steps in step S3, which involve using this model to identify objects in the television screen, include: 1) Obtaining picture frames from a television signal source can be achieved through a software or hardware interface with the television; 2) After acquiring the image, preprocess the image, including adjusting the image size, color correction, and noise reduction. 3) Input the preprocessed image frames into the trained object recognition model, and the model will output the categories of objects identified in the image and their probabilities; 4) Set an appropriate probability threshold, usually 60%-70%. Only items with a recognition probability higher than this threshold will be further processed.

5. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... In step S5, the terminal server performs a query and matching in the product database based on the identified item category. It can use the item category name as a keyword to perform a fuzzy or precise query in the product name or category field of the database, and finally identify the corresponding product.

6. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... In step S6, the associated product information is displayed to the user in an intuitive way, which can be done by displaying the product's image, name, price, and purchase link information on one side of the TV screen or through a pop-up window.

7. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... After step S6, the television also provides user interaction functions, such as allowing users to view more product details, add products to the shopping cart, and select the quantity to purchase via remote control or voice control.

8. The method for automatically identifying items on a television screen and associating them with corresponding goods for sale, as described in claim 1, is characterized in that... The television can also guide users through the purchase process when they decide to buy a product. If the purchase is redirected to an external e-commerce platform, it needs to interface with the e-commerce platform to ensure secure redirection and transmission of user information. For some simple products, the payment process can also be completed directly on the television.

Images