Picture searching method, system and equipment based on 8K set top box and household NAS (Network Attached Storage)

Abstract

The invention provides a picture searching method, system and device based on an 8K set top box and a household NAS, relates to the technical field of set top boxes, and aims to solve the problem that hardware cost is increased by an existing picture searching mode. The method comprises the following steps: the 8K set top box receives a target text input by a user; the target text comprises keywords used for describing the picture; the 8K set top box calls a CLIP model to convert the keyword into a target picture vector, and sends a picture search request containing the target picture vector to a household NAS server; and the household NAS server searches the target picture according to the target picture vector, and sends the target picture to the 8K set top box. According to the method, the strong computing power of the 8K set top box is utilized, interconnection and intercommunication between the 8K set top box and the household NAS server and distributed hardware resource sharing are achieved through the distributed soft bus, and the hardware cost can be reduced.

Description

Technical Field

[0001] The present application relates to the technical field of set-top boxes, and provides a method, system and device for searching images based on an 8K set-top box and a home NAS. Background Art

[0002] With the popularity of smartphone photography, the number of photos people take on a daily basis has increased dramatically. As a centralized, efficient and secure photo storage solution, home network attached storage (NAS) servers have gradually become the choice of many families. The main function of home NAS servers is to provide efficient and secure data storage, but their CPU computing performance is usually limited, and they lack dedicated graphics processing units (GPUs) or neural processing units (NPUs), making it difficult to support complex artificial intelligence reasoning calculations and meet users' urgent needs for efficient retrieval of massive images.

[0003] The existing technology is to integrate a powerful GPU or NPU on a home NAS server to achieve instant image recognition and retrieval. This method can provide a faster image search speed, but it will significantly increase the hardware cost. It is suitable for enterprise-level applications in professional environments, but not friendly to cost-sensitive individual users. Summary of the invention

[0004] The present application provides a method, system and device for image search based on an 8K set-top box and a home NAS, which are used to solve the problem that the existing image search method increases hardware costs.

[0005] In a first aspect, a picture search method based on an 8K set-top box and a home NAS is provided, which is applied to a picture search system based on an 8K set-top box and a home NAS, wherein the picture search system comprises an 8K set-top box and a home NAS server connected via a distributed soft bus; the method comprises: The 8K set-top box receives a target text input by a user; the target text includes keywords for describing a picture; The 8K set-top box calls the CLIP model to convert the keyword into a target image vector, and sends an image search request containing the target image vector to the home NAS server; The home NAS server searches for the target image according to the target image vector, and sends the target image to the 8K set-top box.

[0006] Optionally, before the 8K set-top box receives the target text input by the user, the method further includes: The 8K set-top box deploys the CLIP model and sends a connection request to the home NAS server; The home NAS server verifies the connection request and allows the 8K set-top box to access after the verification is passed; The 8K set-top box sends a login request to the home NAS server; The home NAS server verifies the login request and allows the 8K set-top box to log in after the verification is passed.

[0007] Optionally, the 8K set-top box receives a target text input by a user, including: The 8K set-top box receives a target voice input by a user through a voice remote controller, and converts the target voice into a target text.

[0008] Optionally, the home NAS server searches for a target image according to the target image vector, and sends the target image to the 8K set-top box, including: The home NAS server searches a vector database for multiple image identifiers having a similarity greater than a preset similarity according to the target image vector, and sends a list containing the multiple image identifiers to the 8K set-top box; The 8K set-top box sends the target picture identifier to the home NAS server in response to the user's selection operation on the target picture identifier in the list; The home NAS server searches for the target image from the image database according to the target image identifier, and sends the target image to the 8K set-top box.

[0009] Optionally, before the 8K set-top box receives the target text input by the user, the method further includes: The 8K set-top box sends a picture request to the home NAS server; The home NAS server searches for multiple unencoded pictures according to the picture request, and sends the multiple pictures to the 8K set-top box; The 8K set-top box calls the CLIP model to encode each of the multiple pictures, generates a picture vector corresponding to each picture, and sends the picture vectors corresponding to the multiple pictures to the home NAS server; The home NAS server stores the picture vectors corresponding to the multiple pictures in the vector database.

[0010] Optionally, the target text also includes a time range and a geographical location range; the image search request also includes the time range and the geographical location range; The home NAS server searches for multiple image identifiers having a similarity greater than a preset similarity from a vector database according to the target image vector, and sends a list containing the multiple image identifiers to the 8K set-top box, including: The home NAS server selects a plurality of first picture vectors from a vector database according to the time range and the geographical location range; The home NAS server calculates the similarity between the target image vector and the plurality of first image vectors, and retains a plurality of second image vectors having a similarity greater than a preset similarity; The home NAS server determines multiple picture identifiers corresponding to the multiple second picture vectors, and sends a list including the multiple picture identifiers to the 8K set-top box.

[0011] Optionally, the home NAS server determines a plurality of picture identifiers corresponding to the plurality of second picture vectors, and sends a list including the plurality of picture identifiers to the 8K set-top box, including: The home NAS server determines multiple image identifiers corresponding to the multiple second image vectors, sorts the multiple image identifiers in descending order of similarity, writes the sorted multiple image identifiers into a list, and sends the list to the 8K set-top box.

[0012] Optionally, the 8K set-top box is communicatively connected to a television; after the home NAS server searches for a target image according to the target image vector and sends the target image to the 8K set-top box, the method further includes: The 8K set-top box sends the target image to the TV so that the TV displays the target image.

[0013] In a second aspect, a picture search system based on an 8K set-top box and a home NAS is provided, wherein the picture search system comprises an 8K set-top box and a home NAS server connected via a distributed soft bus; wherein: The 8K set-top box is used to receive a target text input by a user; the target text includes keywords for describing a picture; the CLIP model is called to convert the keyword into a target picture vector, and a picture search request including the target picture vector is sent to the home NAS server; The home NAS server is used to search for a target image according to the target image vector and send the target image to the 8K set-top box.

[0014] In a third aspect, the present application provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the image search method based on an 8K set-top box and a home NAS as described in the first aspect.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The present application provides an image search method based on an 8K set-top box and a home NAS, which is applied to an image search system based on an 8K set-top box and a home NAS, wherein the image search system comprises an 8K set-top box and a home NAS server connected via a distributed soft bus; the method comprises: the 8K set-top box receives a target text input by a user; the target text comprises keywords for describing the image; the 8K set-top box calls a CLIP model to convert the keyword into a target image vector, and sends an image search request containing the target image vector to the home NAS server; the home NAS server searches for a target image according to the target image vector, and sends the target image to the 8K set-top box.

[0016] In this application, the powerful computing power and high-performance GPU / NPU of the 8K set-top box and the efficient storage capacity of the home NAS server are used at the same time, and the interconnection between the 8K set-top box and the home NAS device is realized through the distributed soft bus technology. Compared with the solution of directly integrating GPU / NPU into the home NAS server in the prior art, this application uses the existing 8K set-top box and does not need to modify the home NAS server, reducing the additional hardware cost. In addition, the storage and retrieval of pictures are completed locally, avoiding the network transmission of personal data, thereby avoiding privacy leakage. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0018] Figure 1 A schematic diagram of an application scenario provided for an embodiment of the present application; Figure 2 A flowchart of a method for searching images based on an 8K set-top box and a home NAS provided in an embodiment of the present application; Figure 3 Schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 1 ; Figure 4Schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 2 ; Figure 5 Schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 3 . DETAILED DESCRIPTION

[0019] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme in the embodiment of the present application will be clearly and completely described below in conjunction with the drawings in the embodiment of the present application. Obviously, the described embodiment is only a part of the embodiment of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application. In the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other arbitrarily. In addition, although the logical order is shown in the flow chart, in some cases, the steps shown or described can be performed in an order different from that here.

[0020] Faced with users' urgent need for efficient retrieval of massive images, there are currently two solutions that closely integrate AI technology with practical applications.

[0021] Option 1: Direct integration of high-performance computing.

[0022] Integrating a powerful GPU or NPU on a home NAS server to achieve instant image recognition and retrieval. This solution can provide higher speed and efficiency, but it will significantly increase hardware costs and is not friendly to cost-sensitive individual users.

[0023] Solution 2: Client-server model of distributed architecture.

[0024] Users initiate search requests through the client, and photo storage and AI reasoning tasks are centrally handled by the server. This solution is more cost-friendly and suitable for large-scale deployments and service providers, but since users' search requests and potential personal data need to be transmitted over the network, it is not conducive to privacy protection, and network latency may affect user experience.

[0025] In view of this, the embodiment of the present application provides a picture search method based on 8K set-top box and home NAS, which is applied to a picture search system based on 8K set-top box and home NAS. Figure 1 , is a schematic diagram of an application scenario provided in an embodiment of the present application, or can be understood as a structural diagram of an image search system. The image search system includes an 8K set-top box and a home NAS server.

[0026] Among them, 8K set-top boxes are digital TV set-top boxes that support 8K resolution, which can allow users to watch ultra-high-definition (8K) content on TV through TV signal transmission or Internet streaming transmission. Modern families' 8K set-top boxes are equipped with high-performance CPUs and GPUs, and have powerful artificial intelligence reasoning computing capabilities. Home NAS servers are devices dedicated to storing and managing data, usually used for centralized storage, backup, file sharing, media management, etc., and can be accessed through the network, suitable for home use.

[0027] The NAS server 8K set-top box and the home NAS server are connected through a distributed soft bus. The distributed soft bus refers to a bus communication architecture based on software implementation. It distributes the functions of the traditional hardware bus to different nodes in the network through software to achieve flexible communication and data transmission.

[0028] Combination Figure 2 The embodiment of the present application also provides a method for searching images based on an 8K set-top box and a home NAS. The method steps defined in the process related to the image search method can be Figure 1 The picture search system shown is implemented.

[0029] The following will Figure 2 The specific process shown is explained in detail.

[0030] S201. The 8K set-top box receives the target text input by the user.

[0031] Specifically, the 8K set-top box has a built-in image search application (APK), in which users can enter target text. The target text can only include keywords (such as descriptive text), and keywords are used to describe the content of the picture, such as "zoo". The target text can also include keywords, time range, and geographic location range. The time range is used to limit the shooting time of the picture, and the geographic location range is used to limit the shooting location of the picture, such as "zoo in a certain place in a certain year and month".

[0032] S202, the 8K set-top box calls the CLIP model to convert the keyword into a target image vector, and sends an image search request containing the target image vector to the home NAS server.

[0033] The image search application built into the 8K set-top box pre-deploys the Contrastive Language-Image Pre-training (CLIP) model. The CLIP model can effectively understand the complex relationship between natural language and image content through joint training of large amounts of text and image data, allowing users to quickly locate specific images through simple text descriptions.

[0034] Specifically, after obtaining the target text, the 8K set-top box calls the text encoder of the CLIP model to compare the input target text with the features corresponding to its semantic content, generates a digital vector representation, namely the target image vector, and then sends the image search request containing the target image vector to the home NAS server. If the target text only includes keywords, the image search request only includes the target image vector. If the target text includes keywords, time range and geographic location range, the image search request includes the target image vector, time range and geographic location range.

[0035] S203, the home NAS server searches for the target image according to the target image vector, and sends the target image to the 8K set-top box.

[0036] The home NAS server has pre-built image databases and vector databases. The image database is used to store massive images, such as a large number of photos taken by users. The vector database is used to store the image vector of each image in the image database. The image vector refers to the way of converting an image into a digital vector.

[0037] Specifically, after obtaining the target image vector, the home NAS server can search for at least one similar target image vector from the vector database based on the target image vector, search for the corresponding target image from the image database based on the target image identifier corresponding to the target image vector, and send the target image to the 8K set-top box.

[0038] In the embodiment of the present application, the distributed soft bus between the 8K set-top box and the home NAS server realizes device interconnection and distributed hardware resource sharing. The built-in image search application of the 8K set-top box deploys the CLIP model inference function, and uses its high-performance GPU / NPU to quickly implement text encoding, efficiently utilize the hardware resources of the 8K set-top box, reduce user costs, and protect user privacy and data security.

[0039] In a possible embodiment, before executing S201, the 8K set-top box needs to establish a connection with the home NAS server. The process of establishing the connection is described below.

[0040] The 8K set-top box deploys the CLIP model and sends a connection request to the home NAS server; the home NAS server verifies the connection request and allows the 8K set-top box to access after the verification is passed; the 8K set-top box sends a login request to the home NAS server; the home NAS server verifies the login request and allows the 8K set-top box to log in after the verification is passed.

[0041] Please refer to Figure 3 , which is a schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 1 The 8K set-top box is installed with an image search application (APK) as the client. The home NAS server is installed with an image search application as the server. The connection between the 8K set-top box and the home NAS server actually refers to the connection between the client and the server.

[0042] S301. Client deploys CLIP model.

[0043] Specifically, the 8K set-top box deploys the CLIP model, initializes the inference engine, and loads the CLIP model file.

[0044] S302: Client initialization.

[0045] S303, server initialization.

[0046] S304: The Client sends a connection request to the Server.

[0047] Specifically, a connection request is a request from a Client to attempt to establish a communication connection with a Server.

[0048] S305: The server determines whether to accept the client connection.

[0049] Specifically, the server verifies the connection request, for example, verifies whether the format of the connection request complies with the protocol standard, verifies the IP address based on the firewall and security group rules, etc. If the verification is successful, the client connection is accepted and S306 is continued. If the verification is not successful, the client connection is not accepted and S303 is re-executed.

[0050] S306: The server sends a connection success message to the client.

[0051] S307: The Client sends a login request to the Server.

[0052] Specifically, a login request is a request sent by the Client to the Server for user identity authentication, and the login request contains information such as user name, password, and token.

[0053] S308: The server determines whether to accept the client's login.

[0054] Specifically, the server verifies the login request, such as verifying whether the user name and password are correct, whether the token is expired, etc. If the verification is successful, the client login is accepted and S309 is continued. If the verification is not successful, the client login is not accepted and S303 is re-executed.

[0055] S309: The server sends a login success message to the client.

[0056] In the embodiment of the present application, by dividing it into two independent verification steps, connection request and login request, the home NAS server can more finely control the access process of the 8K set-top box. The connection request ensures that only legitimate devices can access the network, while the login request further verifies the identity of the device, thereby improving security.

[0057] In a possible embodiment, before executing S201, the method further includes: The 8K set-top box sends a picture request to the home NAS server; the home NAS server searches for multiple unencoded pictures according to the picture request, and sends the multiple pictures to the 8K set-top box; the 8K set-top box calls the CLIP model to encode each of the multiple pictures, generates a picture vector corresponding to each picture, and sends the picture vectors corresponding to the multiple pictures to the home NAS server; the home NAS server stores the picture vectors corresponding to the multiple pictures in the vector database.

[0058] Please refer to Figure 4 , which is a schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 2 The 8K set-top box is installed with an image search application (APK) as a client. The home NAS server is installed with an image search application as a server. The interaction between the 8K set-top box and the home NAS server is actually the interaction between the client and the server.

[0059] S401. The Client sends a picture request to the Server.

[0060] Specifically, the image request is used by the Client to request an unencoded image from the Server.

[0061] S402: The server receives a request for an image.

[0062] S403: The server searches for uncoded pictures in the picture database according to the picture request.

[0063] Specifically, after receiving the picture request, the server traverses all pictures stored in the picture database to find out multiple uncoded pictures.

[0064] S404: The server sends the image to the client.

[0065] Specifically, the server returns multiple unencoded pictures to the client.

[0066] S405: The Client receives the image, calls the CLIP model to infer the image, and obtains the image vector.

[0067] Specifically, after receiving the picture returned by the server, the client calls the image encoder of the CLIP model to infer each picture and convert it into a picture vector.

[0068] S406: The Client sends the image vector to the Server.

[0069] S407. The server receives the picture vector.

[0070] S408. The server stores the image vector in a vector database.

[0071] In an embodiment of the present application, the home NAS server is not only responsible for storing massive amounts of images, but also deploys a dedicated vector database for storing the image vector information of each image. Subsequent vector searches based on the vector database can quickly find similar images without having to compare each image one by one. This is more efficient than traditional keyword or content-based searches and can greatly improve search speed.

[0072] There are two specific implementations of S201, which are introduced below respectively.

[0073] In the first method, the 8K set-top box receives the target text entered by the user through the virtual keyboard.

[0074] During the specific implementation process, users can use the remote control of the 8K set-top box to select letters, numbers or symbols in the virtual keyboard one by one, so as to enter the target text in the picture search application of the 8K set-top box.

[0075] In the embodiment of the present application, the virtual keyboard input method has high accuracy but low efficiency, and is suitable for scenarios that require high-precision input and have high requirements for privacy protection.

[0076] The second method is that the 8K set-top box receives the target voice input by the user through the voice remote control and converts the target voice into target text.

[0077] In the specific implementation process, the image search application (APK) of the 8K set-top box can support voice input of target text through a Bluetooth voice remote control, and users can easily initiate voice commands, such as "find photos of cats."

[0078] In the embodiments of the present application, the voice input method is faster and more convenient for users, but it may be affected by accents, noise, etc., and the risk of voice recognition errors is higher. It is suitable for quiet scenes where the user is busy or unable to operate the remote control.

[0079] There are many specific implementations of S203, which are introduced below respectively.

[0080] Method 1: The home NAS server searches for the target image vector with the greatest similarity from the vector database based on the target image vector, searches for the target image from the image database based on the target image identifier corresponding to the target image vector, and sends the target image to the 8K set-top box.

[0081] During the specific implementation process, the home NAS server can use distance measurement methods such as cosine similarity and Euclidean distance to calculate the similarity between the target image vector and each image vector in the vector database, determine the image vector with the greatest similarity as the target image vector, determine the target image identifier corresponding to the target image vector, find the target image corresponding to the target image identifier from the image database, and send the target image to the 8K set-top box.

[0082] In an embodiment of the present application, after the user sends a request, the NAS server automatically searches for and directly sends the most similar target image. The data transmission volume is small, network bandwidth can be saved, and the interaction steps are reduced. The user experience is very smooth and no additional operations are required.

[0083] Method 2: The home NAS server searches for multiple image identifiers with a similarity greater than a preset similarity from the vector database based on the target image vector, and sends a list containing multiple image identifiers to the 8K set-top box; the 8K set-top box responds to the user's selection operation on the target image identifier in the list, and sends the target image identifier to the home NAS server; the home NAS server searches for the target image from the image database based on the target image identifier, and sends the target image to the 8K set-top box.

[0084] In the specific implementation process, the home NAS server can calculate the similarity between the target image vector and each image vector in the vector database, determine the image identifier corresponding to the image vector with a similarity greater than the preset similarity as multiple image identifiers with a similarity greater than the preset similarity, and send a list containing multiple image identifiers to the 8K set-top box. The list can also include a preview image (with a lower resolution) corresponding to each image identifier. The user can select the target image identifier in the list based on the preview image. The home NAS server finds the target image corresponding to the target image identifier from the image database and sends the target image to the 8K set-top box.

[0085] Considering that a single recommendation may not meet personal preferences, in the embodiment of the present application, the home NAS server can filter out multiple similar image identifiers according to the similarity threshold, and the user can more accurately select the target image of interest according to their needs. It not only improves the accuracy and efficiency of image retrieval, but also enhances the personalization and flexibility of user experience. It can provide users with more choice space, reduce the risk of mismatching, improve the fluency of the search and selection process, and meet the needs of different users at the same time.

[0086] In a possible embodiment, the home NAS server selects multiple first image vectors from the vector database according to the time range and the geographic location range; the home NAS server calculates the similarity between the target image vector and the multiple first image vectors, and retains multiple second image vectors with a similarity greater than a preset similarity; the home NAS server determines multiple image identifiers corresponding to the multiple second image vectors, and sends a list containing the multiple image identifiers to the 8K set-top box.

[0087] In the specific implementation process, the home NAS server built a multi-dimensional index structure. In terms of the time dimension, a timeline index was built based on the shooting date and time, supporting fast filtering of pictures by time period. In terms of the geographic dimension, a geospatial index was built based on geographic location information, allowing users to search for pictures by specific location or region. In terms of the content dimension, combined with the text encoding capability of the CLIP model, natural language queries were supported to achieve semantic-level image search.

[0088] In an embodiment of the present application, a user can initiate a multi-condition compound search request. The home NAS server first performs a preliminary screening based on the time range and geographic location range, which can effectively reduce the amount of data and significantly reduce the search time. The screening results are then further searched in combination with the target image vector to ensure the high relevance and accuracy of the search results and improve user satisfaction.

[0089] In a possible embodiment, the home NAS server determines a plurality of picture identifiers corresponding to the plurality of second picture vectors, and sends a list including the plurality of picture identifiers to the 8K set-top box, including: The home NAS server determines multiple image identifiers corresponding to the multiple second image vectors, sorts the multiple image identifiers in descending order of similarity, writes the sorted multiple image identifiers into a list, and sends the list to the 8K set-top box.

[0090] In the embodiment of the present application, through sorting and display, users can see pictures that are closest to their goals or interests, thereby reducing the time and effort of searching and screening.

[0091] In a possible embodiment, the 8K set-top box is communicatively connected to the TV; after executing S203, the method may further include: the 8K set-top box sends the target image to the TV, so that the TV displays the target image.

[0092] In an embodiment of the present application, the searched target image is displayed on a connected large-screen TV in an intuitive manner, and combined with the display advantages of the large-screen TV, an intuitive and convenient multimedia browsing and management interface is provided to the user.

[0093] Please refer to Figure 5 , which is a schematic diagram of the interaction between the 8K set-top box and the home NAS server provided in the embodiment of the present application Figure 3 The 8K set-top box is installed with an image search application (APK) as a client. The home NAS server is installed with an image search application as a server. The interaction between the 8K set-top box and the home NAS server is actually the interaction between the client and the server.

[0094] S501: The Client receives a target text input by a user.

[0095] S502: The Client calls the CLIP model to infer the target text and obtain the target image vector.

[0096] S503: The Client sends an image search request including a target image vector to the Server.

[0097] S504, the server receives the image search request, searches the vector database according to the target image vector, sorts the vectors according to similarity, and obtains a list containing image identifiers.

[0098] S505: The server sends a list containing picture identifiers to the client.

[0099] S506: The Client receives the list including the image identifiers and determines the target image identifier selected by the user.

[0100] S507: The Client sends the target image identifier to the Server.

[0101] S508: The server receives the picture identification request and finds the corresponding target picture according to the target picture identification.

[0102] S509: The server sends the target image to the client.

[0103] S510: The Client receives and displays the target image.

[0104] In summary, the embodiment of the present application provides an image search method based on 8K set-top box and home NAS. By introducing distributed soft bus technology, the advantages of both are fully utilized, namely the powerful computing power and high-performance GPU / NPU of the 8K set-top box, and the efficient storage function of the home NAS server, to achieve real-time collaboration and computing power sharing between hardware resources, creating a new image retrieval experience. Users can enter keywords through the remote control of the set-top box, or use the convenience of the voice remote control to easily explore the massive image database stored in the home NAS server with natural language commands. This method has the following technical advantages: 1. Interconnection between devices: 8K set-top boxes and home NAS servers support the same communication protocol, thus achieving seamless data exchange.

[0105] 2. Distributed hardware resource sharing: Through the task distribution mechanism of the distributed soft bus, complex artificial intelligence tasks are transferred from the home NAS server to the 8K set-top box for execution, making full use of its powerful computing power and NPU and GPU acceleration performance to quickly realize image encoding and text encoding.

[0106] 3. Localized processing: The storage and retrieval of photos are completed locally, avoiding the network transmission of personal data and thus preventing privacy leakage.

[0107] 4. Cost-effectiveness: Compared with directly integrating NPU and GPU into home NAS servers, this application utilizes existing high-performance 8K set-top boxes, reducing additional hardware costs.

[0108] 5. User experience optimization: Combining the display advantages of large-screen TVs, it provides users with an intuitive and convenient multimedia browsing and management interface.

[0109] Based on the same inventive concept, continue to refer to Figure 1 , an embodiment of the present application also provides an image search system based on an 8K set-top box and a home NAS, the image search system comprising an 8K set-top box and a home NAS server connected via a distributed soft bus.

[0110] The 8K set-top box is used to receive a target text input by a user; the target text includes keywords for describing a picture; the CLIP model is called to convert the keywords into a target picture vector, and a picture search request including the target picture vector is sent to the home NAS server; The home NAS server is used to find the target image according to the target image vector and send the target image to the 8K set-top box.

[0111] Optionally, the 8K set-top box is further used to deploy the CLIP model and send a connection request to the home NAS server before the 8K set-top box receives the target text input by the user; The home NAS server is also used to verify the connection request, and after verification, the 8K set-top box is allowed to access; The 8K set-top box is also used to send login requests to the home NAS server; The home NAS server is also used to verify the login request, and the 8K set-top box is allowed to log in after the verification is passed.

[0112] Optionally, the 8K set-top box is specifically used to: receive a target voice input by a user through a voice remote control, and convert the target voice into a target text.

[0113] Optionally, the home NAS server is further used to find multiple image identifiers from a vector database according to the target image vector, and send a list including the multiple image identifiers to the 8K set-top box; The 8K set-top box is further configured to send the target image identifier to the home NAS server in response to a user's selection operation on the target image identifier in the list; The home NAS server is also used to find the target image from the image database according to the target image identifier and send the target image to the 8K set-top box.

[0114] Optionally, the 8K set-top box is further used to send a picture request to the home NAS server before receiving the target text input by the user; The home NAS server is also used to find multiple unencoded pictures based on picture requests and send multiple pictures to the 8K set-top box; The 8K set-top box is also used to call the CLIP model to encode each of the multiple images, generate an image vector corresponding to each image, and send the image vectors corresponding to the multiple images to the home NAS server; The home NAS server is also used to store the image vectors corresponding to the multiple images in a vector database.

[0115] Optionally, the target text also includes a time range and a geographic location range; the image search request also includes a time range and a geographic location range; the home NAS server is specifically used to: filter out multiple first image vectors from a vector database based on the time range and the geographic location range; calculate the similarity between the target image vector and the multiple first image vectors, and retain multiple second image vectors with a similarity greater than a preset similarity; determine multiple image identifiers corresponding to the multiple second image vectors, and send a list containing the multiple image identifiers to the 8K set-top box.

[0116] Optionally, the home NAS server is specifically used to: determine multiple image identifiers corresponding to multiple second image vectors, sort the multiple image identifiers in order of similarity from large to small, write the sorted multiple image identifiers into a list, and send the list to the 8K set-top box.

[0117] Optionally, the 8K set-top box is communicatively connected to the TV; the 8K set-top box is also used to search for the target image according to the target image vector on the home NAS server, and after sending the target image to the 8K set-top box, the target image is sent to the TV so that the TV displays the target image.

[0118] It should be noted that each module in the image search system based on 8K set-top box and home NAS in this embodiment corresponds one by one to each step in the image search method based on 8K set-top box and home NAS in the aforementioned embodiment. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned image search method based on 8K set-top box and home NAS, which will not be repeated here.

[0119] In addition, in one embodiment, the present application also provides a computer device, which includes a processor, a memory, and a computer program stored in the memory, and the computer program implements the aforementioned image search method based on 8K set-top box and home NAS when executed by the processor.

[0120] In addition, in one embodiment, the present application also provides a computer storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the aforementioned image search method based on an 8K set-top box and a home NAS is implemented.

[0121] In some embodiments, the computer readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or various devices including one or any combination of the above memories. The computer may be various computing devices including intelligent terminals and servers.

[0122] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0123] As an example, executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file storing other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0124] By way of example, executable instructions may be deployed to be executed on one computing device, or on multiple computing devices located at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0125] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0126] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not indicate the advantages or disadvantages of the embodiments.

[0127] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a disk, or an optical disk), and includes a number of instructions for a multimedia terminal device to execute the methods described in each embodiment of the present application.

[0128] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A picture search method based on 8K set-top box and home NAS, characterized in that: Applied to an image search system based on an 8K set-top box and a home NAS, the image search system includes an 8K set-top box and a home NAS server connected via a distributed soft bus; the method includes: The 8K set-top box receives a target text input by a user; the target text includes keywords for describing a picture; The 8K set-top box calls the CLIP model to convert the keyword into a target image vector, and sends an image search request containing the target image vector to the home NAS server; The home NAS server searches for the target image according to the target image vector, and sends the target image to the 8K set-top box.

2. The image search method based on 8K set-top box and home NAS according to claim 1, characterized in that: Before the 8K set-top box receives the target text input by the user, the method further includes: The 8K set-top box deploys the CLIP model and sends a connection request to the home NAS server; The home NAS server verifies the connection request and allows the 8K set-top box to access after the verification is passed; The 8K set-top box sends a login request to the home NAS server; The home NAS server verifies the login request and allows the 8K set-top box to log in after the verification is passed.

3. The image search method based on 8K set-top box and home NAS according to claim 1, characterized in that: The 8K set-top box receives a target text input by a user, including: The 8K set-top box receives a target voice input by a user through a voice remote controller, and converts the target voice into a target text.

4. The image search method based on 8K set-top box and home NAS according to claim 1, characterized in that: The home NAS server searches for a target image according to the target image vector, and sends the target image to the 8K set-top box, including: The home NAS server searches a vector database for multiple image identifiers having a similarity greater than a preset similarity according to the target image vector, and sends a list containing the multiple image identifiers to the 8K set-top box; The 8K set-top box sends the target picture identifier to the home NAS server in response to the user's selection operation on the target picture identifier in the list; The home NAS server searches for the target image from the image database according to the target image identifier, and sends the target image to the 8K set-top box.

5. The image search method based on 8K set-top box and home NAS as claimed in claim 4, characterized in that: Before the 8K set-top box receives the target text input by the user, the method further includes: The 8K set-top box sends a picture request to the home NAS server; The home NAS server searches for multiple unencoded pictures according to the picture request, and sends the multiple pictures to the 8K set-top box; The 8K set-top box calls the CLIP model to encode each of the multiple pictures, generates a picture vector corresponding to each picture, and sends the picture vectors corresponding to the multiple pictures to the home NAS server; The home NAS server stores the picture vectors corresponding to the multiple pictures in the vector database.

6. The image search method based on 8K set-top box and home NAS as claimed in claim 4, characterized in that: The target text also includes a time range and a geographical location range; the image search request also includes the time range and the geographical location range; The home NAS server searches for multiple image identifiers having a similarity greater than a preset value from a vector database according to the target image vector, and sends a list containing the multiple image identifiers to the 8K set-top box, including: The home NAS server selects a plurality of first picture vectors from a vector database according to the time range and the geographical location range; The home NAS server calculates the similarity between the target image vector and the plurality of first image vectors, and retains a plurality of second image vectors having a similarity greater than a preset similarity; The home NAS server determines multiple picture identifiers corresponding to the multiple second picture vectors, and sends a list including the multiple picture identifiers to the 8K set-top box.

7. The image search method based on 8K set-top box and home NAS as claimed in claim 6, characterized in that: The home NAS server determines a plurality of picture identifiers corresponding to the plurality of second picture vectors, and sends a list including the plurality of picture identifiers to the 8K set-top box, including: The home NAS server determines multiple image identifiers corresponding to the multiple second image vectors, sorts the multiple image identifiers in descending order of similarity, writes the sorted multiple image identifiers into a list, and sends the list to the 8K set-top box.

8. The image search method based on 8K set-top box and home NAS as claimed in claim 1, characterized in that: The 8K set-top box is communicatively connected with the television; After the home NAS server searches for the target picture according to the target picture vector and sends the target picture to the 8K set-top box, the method further includes: The 8K set-top box sends the target image to the TV so that the TV displays the target image.

9. An image search system based on 8K set-top box and home NAS, characterized in that: The image search system includes an 8K set-top box and a home NAS server connected via a distributed soft bus; wherein: The 8K set-top box is used to receive a target text input by a user; the target text includes keywords for describing a picture; the CLIP model is called to convert the keyword into a target picture vector, and a picture search request including the target picture vector is sent to the home NAS server; The home NAS server is used to search for a target image according to the target image vector and send the target image to the 8K set-top box.

10. A computer device, characterized in that: The computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the image search method based on an 8K set-top box and a home NAS as described in any one of claims 1-7.

Images