Multimedia material publishing method and device and computer equipment

Abstract

The invention relates to a multimedia material publishing method and device, computer equipment, a storage medium and a computer program product. The method can be applied to an application scene in which a vehicle-mounted terminal, a cloud server or other equipment interacts with a multimedia material publishing platform, and comprises the following steps: in response to a material selection operation, obtaining a material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining the number of lenses in the material based on the color histogram; performing color distribution processing on an image in the material to obtain a color distribution value of the image; determining a color entropy value of an image in the material based on the color distribution value; processing the number of the shots and the color entropy value of the image to obtain a material publishing probability; and when the material release probability meets a preset release condition, performing pre-release processing on the material. By adopting the method, the publishing waiting time of the multimedia materials can be effectively reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, and in particular to a multimedia material publishing method and device, a computer device, a storage medium, and a computer program product. BACKGROUND

[0002] With the development of computer technology and Internet technology, the publishing scenarios of multimedia materials are becoming more and more extensive. For example, with the continuous popularity of mobile terminals and the acceleration of network access speed, short videos gradually gain the favor of users due to their short, fast, and high-flow characteristics. People like to perform various personalized editing processes on short videos during the short video shooting process and the short video publishing process, such as adding subtitles, adding icons, and doodling. These editing processes can greatly enrich the video content and meet the personalized needs of users.

[0003] However, in the current multimedia material publishing method, various new media platforms or applications mainly use a switch to control the uploading of multimedia materials to optimize the publishing time consumption of multimedia materials. For example, if the switch for automatically uploading multimedia materials is set to the on state in advance, the background server of various new media platforms or applications will automatically upload the materials authorized by each user to the cloud. However, this processing method will also upload some useless materials, which will increase the storage cost of the cloud. Therefore, how to effectively reduce the publishing waiting time consumption of multimedia materials has become a problem to be solved. SUMMARY

[0004] Therefore, it is necessary to provide a multimedia material publishing method, device, computer device, computer readable storage medium, and computer program product to effectively reduce the publishing waiting time consumption of multimedia materials and effectively reduce the storage cost of multimedia materials in the cloud.

[0005] In a first aspect, the present application provides a multimedia material publishing method. The method comprises: in response to a material selection operation, obtaining a material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining a number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

[0006] In a second aspect, the present application provides a multimedia material publishing device. The device comprises: an obtaining module, configured to obtain a material specified by a material selection operation in response to the material selection operation; a determining module, configured to obtain a color histogram of an image in the material, and determine a number of shots in the material based on the color histogram; a processing module, configured to perform color distribution processing on the image in the material to obtain a color distribution value of the image; the determining module is further configured to determine a color entropy value of the image based on the color distribution value; the processing module is further configured to process the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability satisfies a preset publishing condition, perform pre-publishing processing on the material.

[0007] In an embodiment, the device further comprises: a quantizing module, configured to quantize a hue value domain and a saturation value domain in the color histogram to obtain a quantized hue value domain and a quantized saturation value domain; and the determining module is further configured to determine the number of shots in the material based on the quantized hue value domain and the quantized saturation value domain.

[0008] In an embodiment, the color histogram comprises an HSV histogram; the device further comprises: a dividing module, configured to divide a hue value domain in the HSV histogram to obtain a first number of partition categories corresponding to the hue value domain, and divide a saturation value domain in the HSV histogram to obtain a second number of partition categories corresponding to the saturation value domain; and the determining module is further configured to determine the quantized hue value domain and the quantized saturation value domain based on the first number of partition categories and the second number of partition categories.

[0009] In an embodiment, the material comprises a video; the determining module is further configured to determine a target time interval based on a total duration of the video, determine a candidate frame image based on the target time interval; the processing module is further configured to perform weighted processing on the quantized hue value domain and the quantized saturation value domain corresponding to the candidate frame image to obtain a shot parameter of the candidate frame image; the determining module is further configured to determine a difference value between the shot parameters of each pair of adjacent candidate frame images; and determine the number of shots in the video based on the difference value.

[0010] In an embodiment, the obtaining module is further configured to obtain an initial number of shots in the video; the device further comprises: a comparing module, configured to compare each of the difference values with a preset difference value threshold to obtain a comparison result; and when the comparison result indicates that the difference value is greater than the preset difference value threshold, take a sum of the initial number of shots and a target value as the number of shots in the video.

[0011] In an embodiment, the image in the material includes a target image; the determining module is further configured to determine a shot switching frame in the video based on the difference when the comparison result indicates that the difference is greater than a preset difference threshold, and take the shot switching frame as the target image; and the processing module is further configured to perform color distribution processing on the target image to obtain a color distribution value of the target image.

[0012] In an embodiment, the processing module is further configured to perform mean filtering processing on the target image to obtain the color distribution value of the target image; the determining module is further configured to determine a proportion of different colors in the target image based on the color distribution value of the target image; and determine the color entropy value of the target image in the video according to the proportion.

[0013] In an embodiment, the target image includes a first target image and a second target image; the processing module is further configured to perform color distribution processing on the first target image and the second target image respectively to obtain a first color distribution value of the first target image and a second color distribution value of the second target image; the determining module is further configured to determine a first color entropy value of the first target image based on the first color distribution value, determine a second color entropy value of the second target image based on the second color distribution value, and take a sum value of the first color entropy value and the second color entropy value as the color entropy value.

[0014] In an embodiment, the processing module is further configured to process the number of shots and the color entropy value of the image by a prediction model to obtain a material publishing probability; and perform pre-uploading processing on the material when the material publishing probability is greater than a preset publishing probability threshold.

[0015] In an embodiment, the obtaining module is further configured to obtain index information of a target application program by a user; the apparatus further includes: an ordering module configured to perform correlation ordering on the index information as initial features to obtain ordered initial features; a screening module configured to screen target features from the ordered initial features; and a training module configured to take a material color richness and a number of shots in the target features as training data to train an initial prediction model to obtain the prediction model.

[0016] In an embodiment, the index information includes a video sample material, and the processing module is further configured to perform color distribution processing on an image in the video sample material to obtain a color distribution value of the image in the video sample material; the determining module is further configured to determine a color entropy value of the image in the video sample material based on the color distribution value of the image in the video sample material; and take the color entropy value of the image in the video sample material as the material color richness in the target features.

[0017] In one embodiment, the index information comprises video sample material, the obtaining module is further configured to obtain an HSV histogram of an image in the video sample material; the apparatus further comprises a quantizing module configured to quantize a hue value domain and a saturation value domain in the HSV histogram to obtain a quantized hue value domain and a quantized saturation value domain; the determining module is further configured to determine a shot parameter of the image based on the quantized hue value domain and the quantized saturation value domain; determine a difference between the shot parameters of each adjacent image; and determine the number of shots in the target feature based on the difference.

[0018] In a third aspect, the present application provides a computer device. The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program: in response to a material selection operation, obtaining material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining the number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

[0019] In a fourth aspect, the present application provides a computer readable storage medium. The computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps: in response to a material selection operation, obtaining material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining the number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

[0020] In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the following steps: in response to a material selection operation, obtaining material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining a number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

[0021] The multimedia material publishing method, device, computer device, storage medium and computer program product described above, by responding to a material selection operation, obtaining material specified by the material selection operation; obtaining a color histogram of an image in the material, and determining a number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; and when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material. Since the color entropy value of the image in the material is determined based on the color distribution value obtained by performing color distribution processing on the image in the material, the material publishing probability predicted based on the number of shots in the material and the color entropy value of the image in the material is more accurate, so that the subsequent process of determining whether to perform pre-publishing processing on the material based on the material publishing probability is also more accurate, that is, only when the material publishing probability meets the preset publishing condition, the process of automatically performing pre-publishing processing on the material is triggered, so that the upload proportion of useless materials can be effectively reduced and avoided, the user's uplink bandwidth is saved, the waiting time for publishing multimedia materials is effectively reduced, and the storage cost of multimedia materials in the cloud can also be effectively reduced. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 An application environment diagram of the multimedia material publishing method in one embodiment;

[0023] Figure 2 A flowchart of the multimedia material publishing method in one embodiment;

[0024] Figure 3 An interface display diagram of the product side in one embodiment;

[0025] Figure 4 A flowchart of the shot recognition algorithm in one embodiment;

[0026] Figure 5 Flowchart for calculating mean filter color entropy in one embodiment;

[0027] Figure 6 Diagram for predicting effect of machine learning model in one embodiment;

[0028] Figure 7 Diagram for images with different color richness in one embodiment;

[0029] Figure 8 Diagram for calculating color entropy in traditional way in one embodiment;

[0030] Figure 9 Diagram for defects in traditional algorithm in one embodiment;

[0031] Figure 10 Diagram for predicting effect of method provided by the present application in one embodiment;

[0032] Figure 11 Block diagram of structure of publishing device of multimedia material in one embodiment;

[0033] Figure 12 Internal structure diagram of computer device in one embodiment. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0035] Cloud technology refers to a kind of hosting technology that unifies a series of resources such as hardware, software and network in wide area network or local area network to realize data calculation, storage, processing and sharing.

[0036] Cloud storage is a new concept extended and developed on the basis of cloud computing concept. Distributed cloud storage system (hereinafter referred to as storage system) refers to a storage system that cooperatively works together to provide data storage and business access functions by applying functions such as cluster application, grid technology and distributed storage file system, etc. to collect a large number of storage devices (storage devices are also referred to as storage nodes) of various types in the network through application software or application interface.

[0037] Artificial Intelligence (AI) is the use of digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology of computer science, which tries to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence is the design principle and implementation method of various intelligent machines, so that machines have the functions of perception, reasoning and decision-making.

[0038] With the research and progress of artificial intelligence technology, artificial intelligence technology has been researched and applied in many fields, such as common smart home, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned vehicles, autonomous vehicles, drones, robots, smart medical care, smart customer service, Internet of Vehicles, autonomous driving, intelligent transportation, etc. It is believed that with the development of technology, artificial intelligence technology will be applied in more fields and play an increasingly important role.

[0039] Pre-training model (PTM), also known as cornerstone model or large model, refers to a deep neural network (DNN) with large parameters. It is trained on a large amount of unlabeled data, and the function approximation ability of the large parameter DNN is used to extract common features from the data. Through fine tuning, parameter efficient fine tuning (PEFT), prompt-tuning and other technologies, PTM is suitable for downstream tasks. Therefore, pre-training model can achieve ideal results in few-shot or zero-shot scenarios. PTM can be divided into language models (ELMO, BERT, GPT), visual models (swin-transformer, ViT, V-MOE), speech models (VALL-E), and multi-modal models (ViBERT, CLIP, Flamingo, Gato) according to the data modalities processed, among which multi-modal model refers to a model that establishes feature representation of two or more data modalities. Pre-training model is an important tool for outputting artificial intelligence generated content (AIGC), and can also be used as a general interface connecting multiple specific task models.

[0040] Model parallel computing refers to distributing the computing tasks of a model to multiple computing devices (such as CPU, GPU, TPU, etc.) for simultaneous computation, thereby accelerating the training and inference of the model. Model parallel computing can effectively utilize computing resources, improving the computing efficiency and training speed of the model.

[0041] It should be noted that in the following description, the terms "first", "second", and "third" are merely used to distinguish similar objects, and do not represent a specific order of the objects. Understandably, "first", "second", and "third" can be interchanged in a specific order or sequence as allowed, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein.

[0042] The multimedia material publishing method provided by the embodiments of the present application can be applied to the application environment as shown in the figure. Figure 1 As shown in the figure, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data required to be processed by the server 104. The data storage system can be integrated on the server 104, or placed on the cloud or other network servers. That is, the terminal 102 can interact with the multimedia material publishing platform, i.e. the server 104. The terminal 102 acquires the material specified by the material selection operation in response to the material selection operation. Further, the terminal 102 can acquire the color histogram of the image in the material, and determine the number of shots in the material based on the color histogram. The terminal 102 performs color distribution processing on the image in the material to obtain the color distribution value of the image, and determines the color entropy value of the image in the material based on the color distribution value. Further, the terminal 102 can process the number of shots and the color entropy value of the image to obtain the material publishing probability. When the material publishing probability meets the preset publishing condition, the terminal 102 can perform pre-publishing processing on the material, such as pre-uploading the material to the multimedia material publishing platform, i.e. the server 104, to reduce the waiting time when the material is formally published.

[0043] The terminal 102 can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart television, a smart watch, an Internet of Things device, and a portable wearable device. The Internet of Things device can be a smart vehicle-mounted device. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, etc.

[0044] The server 104 can be a standalone physical server, or a service node in a blockchain system. The service nodes in the blockchain system form a peer-to-peer (Peer To Peer) network among each other. The Peer To Peer protocol is an application layer protocol running on the Transmission Control Protocol (TCP) protocol.

[0045] In addition, the server 104 can also be a server cluster composed of multiple physical servers, and can be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and basic cloud computing services such as big data and artificial intelligence platforms.

[0046] The terminal 102 and the server 104 can be connected through a communication connection mode such as Bluetooth, a universal serial bus (USB), or a network, which is not limited in the present application.

[0047] In one embodiment, as shown in Figure 2 , a multimedia material publishing method is provided, which can be executed by the server or the terminal alone, or by the server and the terminal together. Taking the terminal in Figure 1 as an example, the method includes the following steps:

[0048] In step 202, in response to a material selection operation, the material specified by the material selection operation is obtained.

[0049] The material selection operation refers to a selection operation triggered by the user for a specific material, which is an operation for opening or editing the specific content and related information of the specific material. For example, the material selection operation in the present application can be a trigger operation of the user, which can trigger a trigger event in the terminal, i.e., a screen input event (Input event). The trigger event can include a click event, a touch event, a tap event, a swipe event, etc. That is, the user can perform different trigger operations on the terminal device, for example, the material selection operation in the present application can include a click operation, a swipe operation, a long press operation, an input operation, and a drag operation, etc. The material selection operation in the present application can be an online editing operation, which refers to an operation triggered by the user for jumping to an editing page of a specific material. For example, as shown in Figure 3 , a product-side interface display diagram is shown. The user can select a specific material in the material selection page shown in Figure 3 (1). Assuming that the user clicks the “cloud” video material in the material selection page shown in Figure 3 (1), the operation of clicking the “cloud” video material triggered by the user is the material selection operation.

[0050] The material refers to various multimedia data stored locally by the terminal device used by the user, such as video, picture, multimedia file, animation, music and the like.

[0051] Specifically, the multimedia material publishing method provided by the application can be widely applied to the fields of personalized material creation such as social interaction, game, film and television production, that is, the devices used by different users (operation objects) can interact with the multimedia information platform (or application program). When the user (operation object) wants to publish an image material or a video material containing a specific object, the user can open the multimedia application (Application, APP) on the terminal by triggering operation, and enter the main page of the multimedia application by selecting operation, that is, the user can log in to the multimedia application (such as image editing application, social application) by triggering operation, and further, the user can initiate a multimedia material publishing request in the main page displayed by the multimedia application by triggering operation. For example, in the main page of the social application (video number application) displayed by the terminal, each object (such as a developer) using the social application can view the specific content and related function information in the social main page, and each object using the social application can also trigger a publishing request for a specific material, for example, the operation object A using the social application can click the "material publishing" icon in the main page, so that the terminal responds to the "material publishing" request triggered by the operation object A in the main page of the social application, and obtains all the materials authorized by the operation object A in the local database, and displays them in the material selection page as shown in Figure 3 (1) of FIG. 11; further, the operation object A can select the material to be edited or published in the material selection page as shown in Figure 3 (1) of FIG. 11, then the terminal responds to the material selection operation triggered by the operation object A in the material selection page as shown in Figure 3 (1) of FIG. 11, the terminal can obtain the material selected by the operation object A.

[0052] For example, in the scenario of publishing multimedia materials in a social application, when user A (operation object) wants to publish an image material or a video material containing a specific object, user A can open a specific social application on the terminal by triggering an operation, and enter the main page of the social application by selecting an operation, that is, user A can log in to the social application by triggering an operation, and further, user A can initiate a multimedia material publishing request in the main page displayed by the social application by triggering an operation. For example, in the main page of the social application (video number application) displayed by the terminal, user A can view the specific content and related function information in the social main page. Assuming that user A clicks the "video publishing" icon in the main page, the terminal responds to the "video publishing" request triggered by user A in the main page of the social application, obtains all video materials authorized by the user A in the local database, and displays them in the material selection page shown in FIG. 1 (1); further, user A can select the video material to be edited or published in the video material selection page shown in FIG. 1 (1). For example, assuming that user A clicks the "cloud" video material in the material selection page shown in FIG. 1 (1), indicating that the user A selects the specified video material "cloud" by triggering the above-mentioned click operation, the terminal responds to the above-mentioned material selection operation (click operation) triggered by user A, obtains the video material "cloud" specified by the material selection operation, and displays the obtained video material "cloud" specified by the material selection operation in the material editing page shown in FIG. 1 (2), so that user A can perform some personalized editing operations on the specified video material "cloud", that is, the video material "cloud" to be published can be obtained. That is, the terminal in the present application can display the specific content corresponding to the video material "cloud" specified by the material selection operation in the page shown in FIG. 1 (2). Figure 3 Figure 3 Figure 3 Figure 3 Figure 3

[0053] Step 204: Obtain the color histogram of the image in the material, and determine the number of shots in the material based on the color histogram.

[0054] The color histogram refers to the color histogram corresponding to each image contained in the material. The color histogram describes the proportion of different colors in the entire image, and does not care about the spatial position of each color, that is, it cannot describe the objects or objects in the image. For example, the color histogram in the present application can be an HSV (Hue, Saturation, Value) histogram. HSV is a color space created according to the intuitive characteristics of colors. HSV color model refers to a visible light subset in H, S, V three-dimensional color space, which contains all colors in a certain color domain. ​​​​​

[0055] It is understood that the color histogram in this application includes, but is not limited to, the HSV histogram, and may also be other color histograms, such as the RGB histogram.

[0056] The number of shots refers to the number of shot changes contained in the entire footage. For example, if a video clip A contains 3 scene changes, that is, 3 scene changes with 3 shots, then the number of shots in video clip A is 3.

[0057] Specifically, in response to a media selection operation, after acquiring the media specified in the operation, the terminal can obtain the color histogram of the image in that media and determine the number of shots in that media based on the color histogram. For example, if the media specified in the media selection operation is... Figure 3 (1) If the selected “cloud” video material has a total duration of 30s, the terminal can obtain the HSV histogram of each frame image contained in the “cloud” video material and determine the number of shots in the “cloud” video material based on the HSV histogram of each frame image.

[0058] For example, let's take video footage as an example. Suppose user A... Figure 3 In the material selection page shown in (1), clicking on the video material with the "cloud" cover indicates that user A has selected the specified video material "cloud" through the aforementioned click operation. The terminal then responds to user A's material selection operation (click operation), obtains the video material "cloud" specified by the material selection operation, and displays the obtained video material "cloud" specified by the material selection operation on the screen shown in (1). Figure 4 In the material editing page shown in (2), user A can perform some personalized editing operations on the specified video material "cloud" to obtain the video material "cloud" to be published.

[0059] Furthermore, such as Figure 4 The diagram shows a flowchart of the lens recognition algorithm provided in this application. Assuming the total duration of the video material "clouds" to be released is 30 seconds, containing 30 frames, the terminal determines the target time interval as 10 seconds based on the total duration of the "cloud" video material (30 seconds), and selects frame 1 (A), frame 20 (B), and frame 30 (C) as candidate frame images based on the target time interval of 10 seconds. Further, the terminal can quantize the hue range H and saturation range S in the HSV histogram of each candidate frame image (i.e., image A, image B, and image C) to obtain... Figure 3The terminal can obtain the quantized hue value domain H' and the quantized saturation value domain S' shown in the figure, and further, the terminal can perform weighted processing on the quantized H' and the saturation value domain S' corresponding to each candidate frame image, so as to obtain the lens parameters of each candidate frame image. The terminal can calculate the difference between the lens parameters of each adjacent candidate frame image, that is, the terminal can calculate the difference d1 between the lens parameters of image A and image B, the difference d2 between the lens parameters of image B and image C, and the difference d3 between the lens parameters of image A and image C, and determine the number of lenses in the "cloud" video material based on the obtained differences (d1, d2, d3).

[0060] In step 206, color distribution processing is performed on the images in the material to obtain color distribution values of the images.

[0061] The color distribution processing refers to statistical processing of the spatial positions (distribution conditions) of each color in the image to obtain the statistical color distribution values, that is, the color distribution processing can also be understood as a method for determining the correlation between different colors in the image. For example, the color distribution processing in the present application can be implemented by using mean filtering.

[0062] The color distribution value of the image refers to the color value obtained by statistical processing of the spatial positions (distribution conditions) of each color in the image, that is, in some cases, the color distribution value in the present application can also be referred to as a color value.

[0063] In step 208, the color entropy value of the image in the material is determined based on the color distribution value.

[0064] The color entropy value refers to using the information entropy value of the image color to represent the color feature of the image, which can be understood as reducing the color histogram of the image from multi-dimension to one-dimension, that is, the color entropy is a method for reducing the dimension of the color histogram of the image.

[0065] Specifically, after the terminal obtains the color histogram of the image in the material and determines the number of lenses in the material based on the color histogram, the terminal can perform color distribution processing on the image in the material to obtain the color distribution value of the image, for example, the terminal can perform mean filtering processing on the image in the material to obtain the color distribution value of the image. Further, the terminal can determine the proportion of different colors in the image based on the obtained color distribution value of the image, and determine the color entropy value of the image in the material according to the proportion of different colors in the image.

[0066] For example, taking a video material as an example. Assuming that the terminal determines the color histogram of the video material as Figure 4After the number of shots in the "cloud" video material shown in (1) is D=3, the terminal can determine 3 shot switching frames from the "cloud" video material based on the number of shots, and take the 3 shot switching frames as target images; further, as shown in Figure 3 As shown in (2), the terminal can perform mean filtering processing on the 3 shot switching frames, i.e., the target images, to obtain the color distribution values of the mean filtered target images, and determine the color entropy values of the target images (i.e., the 3 shot switching frames) based on the color distribution values of the mean filtered target images.

[0067] Step 210, processing the number of shots and the color entropy values of the images to obtain a material publishing probability.

[0068] The material publishing probability refers to the probability of the predicted user publishing a specific material, which is used to reflect the willingness of the user to publish the specific material.

[0069] Step 212, when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

[0070] The preset publishing condition refers to a condition for publishing the material that is pre-configured, such as, in the present application, the preset publishing condition can be: judging whether the predicted material publishing probability is greater than a preset publishing probability threshold, if the predicted material publishing probability is greater than the preset publishing probability threshold, it means that the predicted material publishing probability meets the preset publishing condition.

[0071] The pre-publishing processing refers to a series of pre-publishing processing operations on the specified material, such as, in the present application, the pre-publishing processing includes but is not limited to: pre-uploading the material, pre-processing the content of the material, such as, pre-processing the size of the material, or pre-adjusting the watermark information contained in the material, etc.

[0072] Specifically, after the terminal determines the color entropy values of the images in the material based on the color distribution values of the images, the terminal can process the determined number of shots in the material and the color entropy values of the images in the material through a pre-trained prediction model to obtain the predicted publishing probability of the material, i.e., the material publishing probability; further, the terminal can judge whether the predicted publishing probability of the material, i.e., the material publishing probability, meets a preset publishing condition, when the material publishing probability of the material meets the preset publishing condition, the terminal can automatically perform pre-publishing processing on the material. In addition, when the material publishing probability of the material does not meet the preset publishing condition, the terminal will not automatically perform pre-publishing processing on the material, and the terminal can end the processing procedure.

[0073] For example, taking a video material as an example for illustration. Assuming that the terminal determines the color histogram of the video material as shown in (3) based on the color histogram. Figure 4After the number of shots in the "cloud" video material shown in (1) is D=3, the terminal determines 3 shot switching frames from the "cloud" video material based on the number of shots, and determines the color entropy values of the 3 shot switching frame images, and then the terminal can input the number of shots D=3 in the "cloud" video material and the color entropy values of the 3 shot switching frame images into the pre-trained prediction model as input parameters for processing, and the result output by the prediction model is f=96%, i.e. the material release probability of the "cloud" video material is 96%. Further, the terminal can compare the predicted material release probability f=96% of the "cloud" video material with the preset release probability threshold 90%. Since f=96%>90%, the terminal can pre-release the "cloud" video material, for example, the terminal can pre-upload the "cloud" video material to the cloud or the background server to reduce the waiting time when the "cloud" video material is officially released.

[0074] In this embodiment, in response to the material selection operation, the material specified by the material selection operation is obtained; the color histogram of the image in the material is obtained, and the number of shots in the material is determined based on the color histogram; the color distribution of the image in the material is processed to obtain the color distribution value of the image; the color entropy value of the image in the material is determined based on the color distribution value, and the number of shots and the color entropy value of the image are processed to obtain the material release probability; when the material release probability meets the preset release condition, the material is pre-released. Since the color entropy value of the image in the material is determined based on the color distribution value obtained by processing the color distribution of the image in the material, the material release probability predicted based on the number of shots in the material and the color entropy value of the image in the material is more accurate, so that the subsequent process of determining whether to pre-release the material based on the material release probability is also more accurate, i.e. only when the material release probability meets the preset release condition, the process of automatically pre-releasing the material is triggered, so that the upload proportion of useless materials can be effectively reduced and avoided, the user's uplink bandwidth is saved, the release waiting time of the multimedia material is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0075] In one embodiment, the step of determining the number of shots in the material based on the color histogram comprises:

[0076] Quantizing the hue value domain and the saturation value domain in the color histogram to obtain the quantized hue value domain and the quantized saturation value domain;

[0077] Based on the quantized hue value domain and the quantized saturation value domain, the number of shots in the material is determined.

[0078] The hue range refers to the range of hue values ​​in the color histogram. For example, the hue range in this application could be the H value range in the HSV histogram.

[0079] The saturation range refers to the range of saturation values ​​in the color histogram. For example, the saturation range in this application could be the S value range in the HSV histogram.

[0080] Specifically, let's take video footage as an example for explanation. Figure 3 The diagram shown is a flowchart illustrating the lens recognition algorithm provided in this application. Assuming... Figure 4 (2) The total duration of the video material "cloud" shown is 30 seconds, which contains 30 frames of images. The terminal can quantize the hue range H and saturation range S in the HSV histogram of the 30 frames of images contained in the video material "cloud" (or randomly select 10 frames of images from it) to obtain the following result: Figure 4 The diagram shows the quantized hue range H' and quantized saturation range S'. Further, the terminal can perform weighted processing on the quantized H' and saturation range S' corresponding to each frame image to obtain the lens parameters of each frame. The terminal can calculate the difference between the lens parameters of each adjacent pair of frames. For example, the terminal can calculate the difference d1 between the lens parameters of the first frame image A and the second frame image B, and the difference d2 between the lens parameters of the second frame image B and the third frame image C, respectively. Based on all the obtained differences (including d1 and d2), the number of shots in the "cloud" video material is determined.

[0081] Alternatively, the terminal can determine the target time interval as 10 seconds based on the total duration of the "cloud" video footage of 30 seconds, and select frame 1 (A), frame 20 (B), and frame 30 (C) as candidate frame images based on the target time interval of 10 seconds. Furthermore, the terminal can quantize only the hue range H and saturation range S in the HSV histogram of each candidate frame image (i.e., image A, image B, and image C) to obtain, as shown below. Figure 4The terminal can obtain the quantized hue value range H' and the quantized saturation value range S' shown in the figure; further, the terminal can perform weighted processing on the quantized H' and the saturation value range S' corresponding to each candidate frame image, so as to obtain the lens parameters of each candidate frame image. The terminal can calculate the difference between the lens parameters of each adjacent two candidate frame images, i.e., the terminal can calculate the difference d1 between the lens parameters of image A and image B, the difference d2 between the lens parameters of image B and image C, and the difference d3 between the lens parameters of image A and image C, and determine the number of lenses in the "cloud" video material based on the obtained differences (d1, d2, d3). In this way, by taking the number of lenses in the video material as an index that is very helpful for the accuracy of the predicted material release probability, the uploading proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the multimedia material release waiting time is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0082] In one of the embodiments, the color histogram includes an HSV histogram; the step of quantizing the hue value range and the saturation value range in the color histogram to obtain a quantized hue value range and a quantized saturation value range includes:

[0083] The hue value range in the HSV histogram is divided to obtain a first number of partition categories corresponding to the hue value range;

[0084] The saturation value range in the HSV histogram is divided to obtain a second number of partition categories corresponding to the saturation value range;

[0085] The quantized hue value range and the quantized saturation value range are determined based on the first number of partition categories and the second number of partition categories.

[0086] The first number of partition categories and the second number of partition categories are only used to distinguish different numbers of categories divided by the value ranges of different color parameters in the HSV histogram, for example, as shown in the HSV histogram in Figure 4 The H value range in the HSV histogram is divided into 11 partition categories, i.e., 11 value range partitions. Similarly, as shown in the HSV histogram in Figure 4 The S value range in the HSV histogram is divided into 4 partition categories, i.e., 4 value range partitions, so that there are a total of 11*4=44 possibilities, i.e., the terminal can obtain a quantization frequency table with a range of 0-44 integers after quantizing the hue value range and the saturation value range of the HSV histogram as shown in Figure 3

[0087] Specifically, taking the video material as an example. Assuming that the HSV histogram of the video material is as shown in Figure 4 ​The total duration of the video material "cloud" shown in (2) is 30s, which contains 30 frames of images. The terminal can determine the target time interval as 10s based on the total duration 30s of the "cloud" video material, and select the 1st frame image A, the 20th frame image B and the 30th frame image C as candidate frame images based on the target time interval 10s. Further, the terminal can only quantize the hue value domain H and the saturation value domain S in the HSV histogram of each candidate frame image (i.e. image A, image B and image C). For example, when the terminal quantizes the hue value domain H and the saturation value domain S in the HSV histogram of image C, the terminal can divide the H value domain in the HSV histogram of image C to obtain 11 partition categories corresponding to the H value domain, i.e. H'(0-10), as shown in (3). Figure 4 Meanwhile, the terminal can also divide the S value domain in the HSV histogram of image C to obtain 4 partition categories corresponding to the S value domain, i.e. S'(0-3), as shown in (4). Figure 4 Further, the terminal can determine the quantized hue value domain and the quantized saturation value domain based on the first number (11) of partition categories and the second number (4) of partition categories. For example, the terminal can determine a quantization frequency table with a range of 0-44 integers according to the product of the first number (11) of partition categories and the second number (4) of partition categories, i.e. H' * S' = 11 * 4 = 44, which contains the quantized hue value domain and the quantized saturation value domain. In this way, by taking the material color richness as an index that is very helpful to the accuracy of predicting the material release probability, the upload proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the multimedia material release waiting time is effectively reduced, and the storage cost of multimedia materials in the cloud is also effectively reduced.

[0088] In one embodiment, the material includes a video; based on the quantized hue value domain and the quantized saturation value domain, the step of determining the number of shots in the material includes:

[0089] determining a target time interval based on the total duration of the video;

[0090] determining candidate frame images based on the target time interval;

[0091] performing weighted processing on the quantized hue value domain and the quantized saturation value domain corresponding to the candidate frame images to obtain shot parameters of the candidate frame images;

[0092] determining the difference between the shot parameters of each adjacent candidate frame image;

[0093] determining the number of shots in the video based on the difference.

[0094] The candidate frame image refers to one frame image or multiple frame images selected from the video material based on the same time interval.

[0095] Specifically, as shown in the flowchart of the lens recognition algorithm provided in the method of the present application. Figure 4 Assuming that the total duration of the video material "cloud" to be published is 30s, which contains 30 frame images, the terminal determines the target time interval as 10s based on the total duration 30s of the "cloud" video material, and selects the 1st frame image A, the 20th frame image B and the 30th frame image C as the candidate frame images based on the target time interval 10s; further, the terminal can quantize the hue value domain H and the saturation value domain S in the HSV histogram of each candidate frame image (i.e. image A, image B and image C) as shown in Figure 3 , to obtain the quantized hue value domain H' and the quantized saturation value domain S' of each candidate frame image (i.e. image A, image B and image C), that is, for the three frame images of image A, image B and image C, the terminal can calculate the quantized HS values of the three frame images of image A, image B and image C respectively. Assuming that after the terminal quantizes the hue value domain H and the saturation value domain S in the HSV histogram of each candidate frame image (i.e. image A, image B and image C), the obtained quantized HS values corresponding to the three frame images of image A, image B and image C are a, b and c respectively, the terminal can perform weighted processing on the quantized HS values corresponding to each candidate frame image according to the time sequence and the preset weighting coefficients, so as to obtain the lens parameters of each candidate frame image. The terminal can calculate the difference between the lens parameters of each adjacent candidate frame image to determine whether the candidate frame image is a lens switching frame. For example, when the terminal determines whether the frame image C switches the lens, the terminal can first calculate the lens parameter of image C as: 0.6* c + 0.3 * b + 0.1 * a, wherein 0.6, 0.3 and 0.1 are weighting coefficients, which can be preconfigured or dynamically adapted, and no specific limitation is made here.

[0096] In the embodiments of the present application, by taking the material color richness and the number of lenses in the material as indicators that are very helpful to the accuracy of predicting the material publishing probability, the upload proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the multimedia material publishing waiting time is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0097] In one of the embodiments, the step of determining the number of lenses in the video based on the difference value comprises:

[0098] obtaining the initial number of lenses in the video;

[0099] The difference value is compared with a preset difference threshold value to obtain a comparison result.

[0100] When the comparison result indicates that the difference value is greater than the preset difference threshold value, the sum of the initial number of shots and a target value is taken as the number of shots in the video.

[0101] The initial number of shots refers to the initial number of shots, i.e., an initialization value. For example, the initial number of shots in the present application can be 0.

[0102] The preset difference threshold value refers to a threshold value that is configured in advance to measure the size of the difference between the shot parameters of two frames of images. For example, the preset difference threshold value in the present application can be 0.6.

[0103] The comparison result refers to the comparison result between the difference between the shot parameters of each adjacent candidate frame of image and the preset difference threshold value. For example, the comparison result includes a result indicating that the difference is greater than the preset difference threshold value, and a result indicating that the difference is not greater than the preset difference threshold value.

[0104] The target value refers to a preset value. For example, the target value in the present application can be 1.

[0105] Specifically, the material is a video material "cloud" as shown in (2). Figure 3 The terminal can determine whether a candidate frame of image is a shot switching frame by calculating the difference between the shot parameters of each adjacent candidate frame of image. For example, when the terminal determines whether the image C is a shot switching frame, the terminal can calculate the shot parameter of the image C as 0.6 * c + 0.3 * b + 0.1 * a = 2.3, and assume that the terminal calculates the shot parameter of the image B as 1.5. The terminal can calculate the difference d2 between the shot parameters of the image B and the image C as d2 = 2.3 - 1.5 = 0.8, and compare the difference d2 = 0.8 with the preset difference threshold value 0.6 to obtain a comparison result (d2 = 0.8 > 0.6). That is, when the comparison result indicates that the difference d2 is greater than the preset difference threshold value 0.6, it indicates that the image C is a shot switching frame, and the number of shots can be increased by 1. That is, the terminal can take the sum of the initial number of shots s = 0 and the target value 1 as the number of shots g = 1 in the "cloud" video material, so that the number of shots in the "cloud" video material is 1. Thus, the upload proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the waiting time for publishing multimedia materials is effectively reduced, and the storage cost of multimedia materials in the cloud is also effectively reduced.

[0106] In one embodiment, the image in the source material includes a target image; the step of performing color distribution processing on the image in the source material to obtain the color distribution values ​​of the image includes:

[0107] When the comparison result indicates that the difference is greater than the preset difference threshold, the shot switching frame in the video is determined based on the difference, and the shot switching frame is used as the target image;

[0108] Color distribution processing is performed on the target image to obtain the color distribution values ​​of the target image.

[0109] Among them, a shot switching frame refers to a frame image in which a shot switching occurs. For example, if the difference between the shot parameters of the second frame image and the third frame image is greater than a preset difference threshold, it indicates that there is a shot switching between the second frame image and the third frame image, and the third frame image is used as the shot switching frame (image).

[0110] Specifically, taking the materials as an example Figure 3 (2) The video material "clouds" shown in the example is used for illustration. The terminal performs weighted processing on the quantized HS values ​​corresponding to each candidate frame image based on the time series and preset weighting coefficients. After obtaining the lens parameters of each candidate frame image, the terminal can determine whether a candidate frame image is a lens switching frame by calculating the difference between the lens parameters of each adjacent candidate frame image. For example, when the terminal determines whether the image C is a lens switching frame, the terminal can calculate the lens parameters of image C as: 0.6 * c + 0.3 * b + 0.1 * Given a=2.3, assuming the terminal calculates the lens parameter of image B to be 1.5, the terminal can calculate the difference between the lens parameters of image B and image C as d2=2.3-1.5=0.8, and compare this difference d2=0.8 with the preset difference threshold of 0.6 to obtain the comparison result (d2=0.8>0.6). That is, when the comparison result indicates that the difference d2 is greater than the preset difference threshold of 0.6, it means that the image C is a lens switching image. The terminal can then determine the lens switching frame in the video material "cloud" as image C based on the differences, and take the lens switching frame, i.e., image C, as the target image. Furthermore, the terminal can only perform color distribution processing on the selected target image to obtain the color distribution value of the target image. This allows us to determine whether a candidate frame is a shot transition frame by calculating the difference between the lens parameters of each adjacent candidate frame image. Using the number of shots in the material as a very helpful indicator of the accuracy of the predicted material release probability can effectively reduce and avoid the upload ratio of useless material. This saves users' uplink bandwidth, reduces the waiting time for multimedia material release, and also effectively reduces the storage cost of multimedia material in the cloud.

[0111] In one of the embodiments, the color distribution of the target image is processed to obtain the color distribution value of the target image, including:

[0112] The mean filter processing is performed on the target image to obtain the color distribution value of the target image.

[0113] The color entropy value of the image in the material is determined based on the color distribution value, including:

[0114] The proportion of different colors in the target image is determined based on the color distribution value of the target image.

[0115] The color entropy value of the target image in the video is determined according to the proportion.

[0116] Specifically, the material is a video material "cloud" as shown in Figure 5 (2) is taken as an example for illustration. When the comparison result indicates that the difference d2 is greater than the preset difference threshold 0.6, it indicates that the image C is the image of the switching lens, and the terminal can determine the lens switching frame in the video material "cloud" as the image C based on the difference, and take the lens switching frame, i.e., the image C, as the target image. Further, as shown in Figure 5 , the terminal can only perform color distribution processing on the selected target image, i.e., the terminal can perform mean filter processing on the target image, i.e., the image C, to obtain the color distribution value of the image C after mean filter processing, and determine the proportion of different colors in the image C based on the color distribution value of the image C after mean filter processing. The terminal can determine the color entropy value of the image C in the video material "cloud" according to the proportion of different colors in the image C. The calculation method of determining the color entropy value of the image C in the video material "cloud" is shown in the formula in Figure 3 . Thus, by introducing mean filter, the color entropy determined in the technical solution provided by the present application contains the distribution of each color in the image, and the correlation between colors in the image is also included, so that the predicted material release probability based on the number of lenses in the obtained material and the color entropy value of the image in the material is more accurate, so that the subsequent process of determining whether to pre-release the material based on the material release probability is also more accurate, effectively reducing the release waiting time of the multimedia material, and also effectively reducing the storage cost of the multimedia material in the cloud.

[0117] In one of the embodiments, the target image includes a first target image and a second target image; the color distribution of the target image is processed to obtain the color distribution value of the target image, including:

[0118] respectively, to obtain a first color distribution value of the first target image and a second color distribution value of the second target image;

[0119] The color entropy value of the image in the material is determined based on the color distribution value, comprising:

[0120] The first color entropy value of the first target image is determined based on the first color distribution value;

[0121] The second color entropy value of the second target image is determined based on the second color distribution value;

[0122] The sum of the first color entropy value and the second color entropy value is taken as the color entropy value.

[0123] The first target image and the second target image are only used to distinguish different target images, the first color distribution value and the second color distribution value are only used to distinguish color distribution values corresponding to different target images, and the first color entropy value and the second color entropy value are only used to distinguish color entropy values corresponding to different target images.

[0124] Specifically, the material is a video material "cloud" as shown in Figure 5 For example, when the comparison result indicates that the difference values d1 and d2 are both greater than the preset difference threshold 0.6, it indicates that the two images B and C are both images of a switching shot, and the terminal can determine the shot switching frames in the video material "cloud" to be the two images B and C based on the difference values, and take the shot switching frames, i.e., the two images B and C, as target images. Figure 5 As shown in the flowchart of calculating the average filtering color entropy provided by the present application, the terminal can only perform color distribution processing on the selected target images (i.e., images B and C), i.e., the terminal can perform average filtering processing on the target images, i.e., images B and C, to obtain the color distribution values of the images B and C after average filtering processing, which are k1 and k2 respectively, and based on the color distribution values k1 and k2 of the images B and C after average filtering processing, the terminal can determine the color entropy values of the images B and C in the video material "cloud" to be 1.56 and 1.98 respectively according to the proportions of different colors in the images B and C, and take the sum V=1.56+1.98=3.54 of the color entropy value 1.56 of the image B and the color entropy value 1.98 of the image C as the color entropy value of the target images (including images B and C) in the video material "cloud". The calculation method of determining the color entropy values of the images B and C in the video material "cloud" is as shown in Figure 3The formula shown in the formula. Thus, by introducing the mean filter, the color distribution value of each target image can be quickly and accurately obtained, and the color entropy value of each target image is determined based on the color distribution value of each target image, so that the color entropy determined in the technical solution provided by the present application contains the distribution of each color in the image, and the correlation between colors in the image is also included. Therefore, the predicted material release probability based on the number of shots in the material and the color entropy value of the image in the material is more accurate, so that the subsequent process of determining whether to pre-release the material based on the material release probability is also more accurate, effectively reducing the release waiting time of the multimedia material, and also effectively reducing the storage cost of the multimedia material in the cloud.

[0125] In one embodiment, the step of processing the number of shots and the color entropy value of the image to obtain the material release probability includes:

[0126] The number of shots and the color entropy value of the image are processed by the prediction model to obtain the material release probability.

[0127] When the material release probability meets the preset release condition, the material is pre-released, including:

[0128] When the material release probability is greater than the preset release probability threshold, the material is pre-uploaded.

[0129] Specifically, taking the video material as an example. Assuming that the terminal determines the number of shots in the "cloud" video material as D=3 based on the color histogram as shown in (1) of Figure 3 After the terminal determines 3 shot switching frames from the "cloud" video material based on the number of shots and determines the color entropy values of the 3 shot switching frame images, the terminal can input the number of shots D=3 in the "cloud" video material and the color entropy values of the 3 shot switching frame images as input parameters into the pre-trained prediction model for processing, and the result output by the prediction model is f=96%, i.e. the material release probability of the "cloud" video material is 96%. Further, the terminal can compare the predicted material release probability f=96% of the "cloud" video material with the preset release probability threshold 90%. Since f=96%>90%, the terminal can pre-release the "cloud" video material, for example, the terminal can pre-upload the "cloud" video material to the cloud or the background server to reduce the waiting time when the "cloud" video material is officially released. Thus, the upload proportion of useless materials can be effectively reduced and avoided, the user's uplink bandwidth is saved, the release waiting time of the multimedia material is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0130] In an embodiment, the method further comprises:

[0131] obtaining index information of the user in the target application program;

[0132] performing relevance sorting on the index information as initial features to obtain sorted initial features;

[0133] selecting target features from the sorted initial features;

[0134] taking the material color richness and the number of shots in the target features as training data to train the initial prediction model to obtain a prediction model.

[0135] The target application program refers to a specific application program. For example, the target application program in the present application includes but is not limited to a social application program and a video number application program.

[0136] The index information refers to authorized index information of the user. For example, the index information in the present application includes but is not limited to 50 useful index information such as user age, gender, use frequency, published edited content of the video number (target application), added text at the time of publishing, whether real name, function entry, publishing with expansion link, authentication account, real name authentication, video number fan number, total number of published dynamics, total number of likes, care mode, reselected material, real name authentication information of the application program, watermark detection, etc.

[0137] The relevance sorting refers to sorting from high to low correlation between the initial features and the publishing intention of the user, i.e., sorting according to the correlation (from high to low order) between each initial feature and the publishing intention of the user.

[0138] The initial prediction model refers to an untrained initial model. For example, the initial prediction model in the present application includes but is not limited to a linear regression model, a decision tree model, or a support vector machine model, etc.

[0139] Specifically, the terminal can obtain the index information of different users in the target application program, and perform correlation ranking on the obtained index information as initial features to obtain ranked initial features. Further, the terminal can screen target features from the ranked initial features, and take the material color richness and the number of shots in the target features as training data to train the initial prediction model, so as to obtain the trained prediction model. When applying the prediction model, the terminal can call the prediction model, and process the number of shots in the material and the color entropy value of the target image (shot switching frame image) through the prediction model, so as to quickly and accurately output the predicted publishing probability of different materials. Thus, the upload proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the multimedia material publishing waiting time is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0140] In one embodiment, the index information includes video sample materials, and the method further includes:

[0141] performing color distribution processing on the images in the video sample materials to obtain color distribution values of the images in the video sample materials;

[0142] determining color entropy values of the images in the video sample materials based on the color distribution values of the images in the video sample materials;

[0143] taking the color entropy values of the images in the video sample materials as the material color richness in the target features.

[0144] Specifically, after the terminal screens the target features from the ranked initial features, the terminal can take the material color richness and the number of shots in the target features as training data. That is, the terminal can perform color distribution processing on the images in the video sample materials in the collected index information in advance to obtain color distribution values of the images in the video sample materials. Then, the terminal determines color entropy values of the images in the video sample materials based on the color distribution values of the images in the video sample materials, and takes the color entropy values of the images in the video sample materials as the material color richness (feature) in the target features. Thus, by performing color distribution processing on the images in the material, the color distribution values of the target images can be quickly and accurately obtained, and the color entropy values of the target images are determined based on the color distribution values of the target images. The color entropy determined in the technical solution provided in the present application contains the distribution of each color in the image, and the correlation between colors in the image is also contained. Therefore, the predicted material publishing probability based on the number of shots in the material and the color entropy values of the images in the material is more accurate, so that the process of determining whether to perform pre-publishing processing on the material based on the material publishing probability is also more accurate.

[0145] In an embodiment, the index information comprises video sample material, and the method further comprises:

[0146] obtaining an HSV histogram of an image in the video sample material;

[0147] quantizing a hue value domain and a saturation value domain in the HSV histogram to obtain a quantized hue value domain and a quantized saturation value domain;

[0148] determining a shot parameter of the image based on the quantized hue value domain and the quantized saturation value domain;

[0149] determining a difference between the shot parameters of each adjacent image;

[0150] determining a number of shots in the target feature based on the difference.

[0151] Specifically, after the terminal screens the target feature from the sorted initial features, the terminal can take the material color richness and the number of shots in the target feature as training data, that is, the terminal can obtain the HSV histogram of the image in the video sample material in the index information, and quantize the hue value domain and the saturation value domain in the HSV histogram to obtain the quantized hue value domain and the quantized saturation value domain. The terminal can determine the shot parameter of the image based on the quantized hue value domain and the quantized saturation value domain. Further, the terminal can determine the difference between the shot parameters of each adjacent image, determine the number of shots in the video sample material based on the difference, and take the determined number of shots in the video sample material as the number of shots (feature) in the target feature. Thus, by taking the number of shots in the material as an index that is very helpful for the accuracy of the predicted material publishing probability, the upload proportion of useless materials can be effectively reduced and avoided, the user uplink bandwidth is saved, the multimedia material publishing waiting time is effectively reduced, and the storage cost of the multimedia material in the cloud is also effectively reduced.

[0152] The application also provides an application scenario applying the multimedia material publishing method. Specifically, the multimedia material publishing method is applied in the application scenario as follows:

[0153] In the process of user interaction with a multimedia information (or image editing) platform, the multimedia material publishing method described above can be used. When the user (operation object) wants to generate an image or video containing a specific object, the user can open the multimedia application on the terminal by triggering the operation, and enter the main page of the multimedia application by selecting the operation. Further, the user can initiate a material publishing request by triggering the operation in the main page displayed by the multimedia application, and the terminal responds to the material publishing request triggered by the user to obtain all the materials authorized by the user in the local database and display them in theFigure 3 The material selection page shown in (1) in the foregoing embodiment; further, the user can select the material that the user wants to edit or publish in the material selection page shown in (1) in the foregoing embodiment, and the terminal responds to the material selection operation triggered by the user in the material selection page shown in (1) in the foregoing embodiment. Figure 3 The material selection page shown in (1) in the foregoing embodiment; further, the user can select the material that the user wants to edit or publish in the material selection page shown in (1) in the foregoing embodiment, and the terminal responds to the material selection operation triggered by the user in the material selection page shown in (1) in the foregoing embodiment. Figure 3 The material selection page shown in (1) in the foregoing embodiment; further, the user can select the material that the user wants to edit or publish in the material selection page shown in (1) in the foregoing embodiment, and the terminal responds to the material selection operation triggered by the user in the material selection page shown in (1) in the foregoing embodiment. Thus, when different users interact with the multimedia information platform, the multimedia material publishing method provided in the present application predicts the willingness of the user to click on the publish specified material in the publishing page through the user local data, and when it is predicted that the user will publish the material, the terminal automatically performs the pre-upload processing on the material, thereby achieving the technical effects of reducing the waiting time for material publishing and reducing the cloud storage cost.

[0154] The method provided in the embodiments of the present application can be applied to various multimedia material publishing scenarios. The following takes the scenario in which the user interacts with the social application as an example to describe the multimedia material publishing method provided in the embodiments of the present application.

[0155] In the embodiments of the present application, the precision = the number of times of hitting the CDN pre-upload when publishing / the total number of times of publishing, and the recall rate = the number of times of hitting the CDN pre-upload when publishing / the total number of times of CDN pre-upload.

[0156] With the continuous popularity of mobile terminals and the acceleration of network access speed, short videos gradually gain the favor of users due to their short, fast, and large flow characteristics. In the traditional multimedia material publishing mode, such as the time-consuming optimization for material publishing in various new media platforms or application programs: the upload is performed in advance in the publishing page (controlled by the user switch), and the switch is in the default on state. However, this automatic upload material mode controlled by the switch has a relatively large proportion of useless upload, which wastes the user uplink bandwidth and also wastes the CDN (Content Delivery Network, i.e., content distribution network) storage cost. Therefore, how to effectively reduce the waiting time for multimedia material publishing and also effectively reduce the CDN storage cost becomes a problem to be solved.

[0157] The technical scheme provided by the application can solve these problems at the same time through innovative complete process design:

[0158] According to a series of operation behaviors of the user and historical publishing behaviors, a machine learning scheme is used to predict a probability that the user will publish the material this time on the publishing page. The terminal automatically performs the pre-upload material behavior only when it is predicted that the user will publish. That is, through the user local data and the data of the past publishing behaviors, the machine learning model is trained, and the accuracy rate of more than 90% can be achieved, so as to avoid wasting the bandwidth of the user and save the CDN storage cost.

[0159] On the product side, in the traditional video number publishing business, after the user selects the material, the

CDN upload

[0160] As shown in Figure 3 , the product side style of the technical scheme provided by the application can be as shown in Figure 3 , as shown in Figure 6 (1), when the user selects the specified material, the terminal automatically executes the multimedia material publishing method provided in the embodiment of the application to confirm whether the pre-publishing processing of the material selected by the user is needed, that is, whether the processing flow of pre-uploading the material is triggered.

[0161] On the technology side, as shown in Figure 6 , it is a schematic diagram of the prediction effect of the machine learning model.

[0162] 1. Clear features

[0163] The clear user available information includes but is not limited to: user age, gender, use frequency, video number publishing editing content, publishing text addition, whether real name, function entry, publishing with expansion link, authentication account, real name authentication, video number fan number, total number of published dynamics, total number of likes, care mode, reselected material, real name authentication information, watermark detection, and more than 50 useful indexes.

[0164] 2. Machine learning model prediction

[0165] (1) Linear regression model

[0166] (2) Decision Tree Model

[0167] (3) Support Vector Machine Model

[0168] As shown in Figure 6 , the scheme provided by the embodiments of the present application attempts to use three initial models, and it can be found that the effects of the models on each other are not very different, and in particular, the effects of the decision tree and the support vector machine model are basically equivalent.

[0169] Secondly, it can be determined that the three models have better effects than manually finding constraints, but there is still a gap from the target shown in Figure 6 , and therefore, in order to further improve the model effect, the model needs to be further deepened or the features need to be optimized.

[0170] In the process of training the initial prediction model as shown in Figure 7 , it is found that a good feature has a greater additive effect on the prediction effect, so feature optimization may be the most efficient breakthrough.

[0171] 3. Feature optimization

[0172] The terminal can use the RFE algorithm to filter the feature weight, sort the features collected in the first step, and filter out which features are useful and which features are useless. In addition to mining the relevant information of the user himself, the technical scheme provided in the present application also finds two indexes that are very helpful to the accuracy of predicting whether the user will publish the material: material color richness and video shot number.

[0173] (1) Material color richness

[0174] As shown in Figure 7 , it is a schematic diagram of images with different color richness. For example, from the sense, the image in Figure 7 (1), that is, the left picture, is less rich in color than the image in Figure 7 (2), that is, the right picture. In the traditional way, the indexes for calculating the color richness of an image are color entropy and spatial clustering, but in essence, they are all statistics of the proportion of different colors in an image. For example, since the right image in Figure 8 is richer in color than the left image, the color entropy of the left image calculated by the traditional method is 1, and the color entropy of the right image is 2.32. That is, the disadvantage of the traditional color entropy calculation of color richness is that only color is considered, and the concept of distribution is discarded. For example, as shown in Figure 8 , it is a schematic diagram of calculating color entropy in the traditional way. That is, as shown in Figure 5The color entropy of the second and third blocks shown in the middle is consistent, but the third block is obviously more diverse. The color entropy ignores the distribution problem because it ignores the correlation between colors, so the technical solution provided in the application introduces mean filtering. As shown in Figure 9 The technical solution provided in the embodiments of the application calculates the color entropy by filtering the color mean, and the color entropy of the left part is 1.56, and the color entropy of the right part is 1.98, which well completes the distinction.

[0175] (2) Number of video shots

[0176] There are currently two methods in the industry to count the number of shots: the first is machine learning detection, and the second is the RGB histogram method.

[0177] The industry focus of machine learning is to detect objects in the shot, not the number of shots, so the technical solution provided in the application optimizes the slightly simple traditional algorithm.

[0178] As shown in Figure 9 , it is a schematic diagram of the defects existing in the traditional algorithm. That is, as shown in Figure 10 , the traditional RGB shot recognition method mainly checks by calculating the difference between the RGB histograms of two pictures. After comparing the video samples, it is found that the accuracy rate of falling within the prediction interval is only 64%. The problems existing in this traditional method include: first, the RGB histogram cannot eliminate the brightness index, but the shot switching is often unrelated to brightness, so the HSV histogram used in the embodiments of the application has better effect on brightness processing. Second, the scheme of only comparing two frames is not friendly to gradual transition shots, so the verification frame used in the embodiments of the application should have memory capability.

[0179] The traditional RGB shot recognition algorithm is improved. First, the first point is to replace RGB with HSV, which screens out the influence of brightness on shot switching by fixing the V value, which is the brightness value. Because the value range of H and S is very wide, in order to better quantitatively analyze, it is necessary to segment the value range to obtain a quantization frequency table with an integer range of 0-44. Next, the technical solution provided in the application does not directly calculate the difference between two frames, but adds the time parameter to weight the quantized values according to the time sequence, in order to be compatible with the scene of gradual transition shots. That is, through the way adopted in the technical solution of the application, a shot recognition algorithm with an accuracy of 86% is finally obtained. As shown in Figure 11 , it is a prediction effect diagram of the method provided in the application. In the place where the shot switching is recognized, the color richness is calculated, and through these two indicators, the accuracy and recall rate of the prediction model are successfully improved to more than 90%, achieving the set target.

[0180] It can be understood that the technical solutions provided in the embodiments of the present application can also use a deep learning scheme in the model part.

[0181] The beneficial effects of the technical solutions of the present application include:

[0182] By taking the material color richness and the number of shots in the material as indicators that are very helpful for predicting the accuracy of the material publishing probability, the prediction accuracy in the technical solutions of the present application reaches 95%, effectively saving CDN costs of hundreds of thousands of yuan (desensitization), and effectively reducing the time consumption of material publishing.

[0183] It should be understood that, although each step in the flowchart involved in each embodiment as described above is displayed in sequence according to the arrow, these steps are not necessarily executed in sequence according to the order of the arrow. Unless explicitly stated herein, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least part of the steps in the flowchart involved in each embodiment as described above can include multiple steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least part of other steps or steps or stages in other steps.

[0184] Based on the same inventive concept, the embodiments of the present application also provide a multimedia material publishing device for implementing the multimedia material publishing method described above. The implementation scheme for solving problems provided by the device is similar to the implementation scheme described in the above method, so the specific limitations in one or more multimedia material publishing device embodiments provided below can refer to the limitations of the multimedia material publishing method in the above text, and will not be repeated here.

[0185] In one embodiment, as shown in Figure 12 A multimedia material publishing device is provided, comprising: an acquisition module 1102, a determination module 1104, and a processing module 1106, wherein:

[0186] The acquisition module 1102 is configured to acquire a material specified by a material selection operation in response to the material selection operation;

[0187] The determination module 1104 is configured to acquire a color histogram of an image in the material, and determine the number of shots in the material based on the color histogram;

[0188] The processing module 1106 is configured to perform color distribution processing on the image in the material to obtain a color distribution value of the image;

[0189] The determining module 1104 is further configured to determine a color entropy value of an image in the material based on the color distribution value.

[0190] The processing module 1106 is further configured to process the number of shots and the color entropy value of the image to obtain a material publishing probability, and perform pre-publishing processing on the material when the material publishing probability meets a preset publishing condition.

[0191] In an embodiment, the apparatus further includes a quantizing module configured to quantize a hue value domain and a saturation value domain in the color histogram to obtain a quantized hue value domain and a quantized saturation value domain; and the determining module is further configured to determine the number of shots in the material based on the quantized hue value domain and the quantized saturation value domain.

[0192] In an embodiment, the color histogram includes an HSV histogram; the apparatus further includes a dividing module configured to divide a hue value domain in the HSV histogram to obtain a first number of partition categories corresponding to the hue value domain, and divide a saturation value domain in the HSV histogram to obtain a second number of partition categories corresponding to the saturation value domain; and the determining module is further configured to determine the quantized hue value domain and the quantized saturation value domain based on the first number of partition categories and the second number of partition categories.

[0193] In an embodiment, the material includes a video; the determining module is further configured to determine a target time interval based on a total duration of the video, and determine a candidate frame image based on the target time interval; the processing module is further configured to perform weighted processing on the quantized hue value domain and the quantized saturation value domain corresponding to the candidate frame image to obtain a shot parameter of the candidate frame image; the determining module is further configured to determine a difference value between the shot parameters of each pair of adjacent candidate frame images; and determine the number of shots in the video based on the difference value.

[0194] In an embodiment, the obtaining module is further configured to obtain an initial number of shots in the video; the apparatus further includes a comparing module configured to compare each of the difference values with a preset difference value threshold to obtain a comparison result, and when the comparison result indicates that the difference value is greater than the preset difference value threshold, take a sum of the initial number of shots and a target value as the number of shots in the video.

[0195] In an embodiment, the image in the material includes a target image; the determining module is further configured to determine a shot switching frame in the video based on the difference when the comparison result indicates that the difference is greater than a preset difference threshold, and take the shot switching frame as the target image; and the processing module is further configured to perform color distribution processing on the target image to obtain a color distribution value of the target image.

[0196] In an embodiment, the processing module is further configured to perform mean filtering processing on the target image to obtain the color distribution value of the target image; the determining module is further configured to determine a proportion of different colors in the target image based on the color distribution value of the target image; and determine the color entropy value of the target image in the video according to the proportion.

[0197] In an embodiment, the target image includes a first target image and a second target image; the processing module is further configured to perform color distribution processing on the first target image and the second target image respectively to obtain a first color distribution value of the first target image and a second color distribution value of the second target image; the determining module is further configured to determine a first color entropy value of the first target image based on the first color distribution value; determine a second color entropy value of the second target image based on the second color distribution value; and take a sum value of the first color entropy value and the second color entropy value as the color entropy value.

[0198] In an embodiment, the processing module is further configured to process the number of shots and the color entropy value of the image by a prediction model to obtain a material publishing probability; and perform pre-uploading processing on the material when the material publishing probability is greater than a preset publishing probability threshold.

[0199] In an embodiment, the obtaining module is further configured to obtain index information of a target application program by a user; the apparatus further includes: an ordering module configured to perform correlation ordering on the index information as initial features to obtain ordered initial features; a screening module configured to screen target features from the ordered initial features; and a training module configured to take a material color richness and a number of shots in the target features as training data to train an initial prediction model to obtain the prediction model.

[0200] In an embodiment, the index information includes a video sample material, and the processing module is further configured to perform color distribution processing on an image in the video sample material to obtain a color distribution value of the image in the video sample material; the determining module is further configured to determine a color entropy value of the image in the video sample material based on the color distribution value of the image in the video sample material; and take the color entropy value of the image in the video sample material as the material color richness in the target features.

[0201] In one embodiment, the indicator information includes video sample material, and the acquisition module is further configured to acquire the HSV histogram of the image in the video sample material; the device further includes: a quantization module, configured to quantize the hue range and saturation range in the HSV histogram to obtain the quantized hue range and the quantized saturation range; the determination module is further configured to determine the lens parameters of the image based on the quantized hue range and the quantized saturation range; determine the difference between the lens parameters of each adjacent image; and determine the number of shots in the target feature based on the difference.

[0202] Each module in the aforementioned multimedia material publishing device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0203] In one embodiment, a computer device is provided, which may be a terminal or a server. In this embodiment, the computer device is described as a terminal, and its internal structure diagram is as follows. Figure 12 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for publishing multimedia materials. The display unit of the computer device is used to form a visually visible image. It can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0204] Those skilled in the art will understand that ​The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0205] In an embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, and the processor implementing the steps in the above method embodiments when executing the computer program.

[0206] In an embodiment, a computer readable storage medium is provided, storing a computer program, and the computer program implementing the steps in the above method embodiments when executed by a processor.

[0207] In an embodiment, a computer program product is provided, including a computer program, and the computer program implementing the steps in the above method embodiments when executed by a processor.

[0208] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0209] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0210] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0211] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method of distributing multimedia material, characterized by The method comprises: in response to a material selection operation, obtaining material specified by the material selection operation; acquiring a color histogram of an image in the material, and determining a number of shots in the material based on the color histogram; performing color distribution processing on the image in the material to obtain a color distribution value of the image; determining a color entropy value of the image in the material based on the color distribution value; processing the number of shots and the color entropy value of the image to obtain a material publishing probability; when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material.

2. The method of claim 1, wherein, The number of shots in the material is determined based on the color histogram, comprising: quantizing a hue value domain and a saturation value domain in the color histogram to obtain a quantized hue value domain and a quantized saturation value domain; determining the number of shots in the material based on the quantized hue value domain and the quantized saturation value domain.

3. The method of claim 2, wherein, The color histogram comprises an HSV histogram; the hue value domain and the saturation value domain in the color histogram are quantized to obtain the quantized hue value domain and the quantized saturation value domain, comprising: dividing the hue value domain in the HSV histogram to obtain a first number of partition categories corresponding to the hue value domain; dividing the saturation value domain in the HSV histogram to obtain a second number of partition categories corresponding to the saturation value domain; determining the quantized hue value domain and the quantized saturation value domain based on the first number of partition categories and the second number of partition categories.

4. The method of claim 2, wherein, The material comprises a video; the number of shots in the material is determined based on the quantized hue value domain and the quantized saturation value domain, comprising: determining a target time interval based on a total duration of the video; determining a candidate frame image based on the target time interval; performing weighted processing on the quantized hue value domain and the quantized saturation value domain corresponding to the candidate frame image to obtain a shot parameter of the candidate frame image; determining a difference value between the shot parameters of each adjacent candidate frame image; determining the number of shots in the video based on the difference value.

5. The method of claim 4, wherein, The number of shots in the video is determined based on the difference value, comprising: obtaining an initial number of shots in the video; comparing each difference value with a preset difference value threshold to obtain a comparison result; when the comparison result indicates that the difference value is greater than the preset difference value threshold, taking a sum of the initial number of shots and a target value as the number of shots in the video.

6. The method of claim 5, wherein, The image in the material comprises a target image; the color distribution processing is performed on the image in the material to obtain the color distribution value of the image, comprising: when the comparison result indicates that the difference value is greater than the preset difference value threshold, determining a shot switching frame in the video based on the difference value, and taking the shot switching frame as the target image; performing color distribution processing on the target image to obtain a color distribution value of the target image.

7. The method of claim 6, wherein, The color distribution processing is performed on the target image to obtain the color distribution value of the target image, comprising: perform mean filtering processing on the target image to obtain a color distribution value of the target image; the determining of the color entropy value of the image in the material based on the color distribution value comprises: determining a proportion of different colors in the target image based on the color distribution value of the target image; determining the color entropy value of the target image in the video according to the proportion.

8. The method of claim 6, wherein, The target image includes a first target image and a second target image; the color distribution processing on the target image to obtain the color distribution value of the target image comprises: performing color distribution processing on the first target image and the second target image respectively to obtain a first color distribution value of the first target image and a second color distribution value of the second target image; the determining of the color entropy value of the image in the material based on the color distribution value comprises: determining a first color entropy value of the first target image based on the first color distribution value; determining a second color entropy value of the second target image based on the second color distribution value; taking a sum value of the first color entropy value and the second color entropy value as the color entropy value.

9. The method of claim 1, wherein, The processing of the number of shots and the color entropy value of the image to obtain a material publishing probability comprises: processing the number of shots and the color entropy value of the image by a prediction model to obtain a material publishing probability; when the material publishing probability meets a preset publishing condition, performing pre-publishing processing on the material, comprising: when the material publishing probability is greater than a preset publishing probability threshold, performing pre-uploading processing on the material.

10. The method of claim 9, wherein, The method further comprises: obtaining index information of a target application program by a user; performing correlation sorting on the index information as initial features to obtain sorted initial features; screening target features from the sorted initial features; taking a material color richness and a number of shots in the target features as training data to train an initial prediction model to obtain the prediction model.

11. The method of claim 10, wherein, The index information includes a video sample material, and the method further comprises: performing color distribution processing on images in the video sample material to obtain color distribution values of the images in the video sample material; determining color entropy values of the images in the video sample material based on the color distribution values of the images in the video sample material; taking the color entropy values of the images in the video sample material as the material color richness in the target features.

12. The method of claim 10, wherein, The index information includes a video sample material, and the method further comprises: obtaining an HSV histogram of an image in the video sample material; quantizing a hue value domain and a saturation value domain in the HSV histogram to obtain a quantized hue value domain and a quantized saturation value domain; determining a shot parameter of the image based on the quantized hue value domain and the quantized saturation value domain; determining a difference value between the shot parameters of each adjacent image; determining the number of shots in the target features based on the difference value.

13. A distribution apparatus of multimedia material, characterized by comprising: The device comprises: an acquisition module, configured to acquire a material specified by a material selection operation in response to the material selection operation; The determining module is configured to acquire a color histogram of an image in the material, and determine a number of shots in the material based on the color histogram; The processing module is configured to perform color distribution processing on the image in the material to obtain a color distribution value of the image; The determining module is further configured to determine a color entropy value of the image in the material based on the color distribution value; The processing module is further configured to process the number of shots and the color entropy value of the image to obtain a material publishing probability, and perform pre-publishing processing on the material when the material publishing probability satisfies a preset publishing condition.

14. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to implement the steps of the method in any one of claims 1 to 12.

15. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 12.

16. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 12. The computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 12.

Images