WO2020006962A1

WO2020006962A1 - Method and device for processing picture

Info

Publication number: WO2020006962A1
Application number: PCT/CN2018/116336
Authority: WO
Inventors: 徐珍琦; 王长虎
Original assignee: 北京字节跳动网络技术有限公司
Priority date: 2018-07-03
Filing date: 2018-11-20
Publication date: 2020-01-09
Also published as: CN108921792A; CN108921792B

Abstract

Disclosed in embodiments of the present application are a method and device for processing a picture. One specific implementation of the method comprises: obtaining a picture to be processed; selecting, from a preset standard size set, a standard size corresponding to the picture to be processed as a target size; adjusting the size of the picture to the processed to the target size to obtain an adjusted picture to be processed; and inputting the adjusted picture to be processed to a pre-trained picture processing model to obtain a processing result. The implementation facilitates improving the flexibility of picture size adjustment.

Description

Method and device for processing pictures

This patent application claims the priority of a Chinese patent application filed on July 3, 2018, with application number 201810714637.5, the applicant being Beijing BYTE Network Technology Co., Ltd., and the invention name "Method and Device for Processing Pictures" , The entire application of which is incorporated herein by reference.

Technical field

Embodiments of the present application relate to the field of computer technology, and in particular, to a method and an apparatus for processing a picture.

Background technique

With the development of artificial neural networks, their applications in the field of image processing are becoming more and more widespread. For example, in the fields of picture recognition, picture classification, etc., corresponding neural network models can be used to process pictures.

In the prior art, some processes using neural network models to process pictures usually first scale the size of the picture to be processed to a standard size, and then process the picture. In addition, when the number of pictures to be processed is multiple, the size of each picture to be processed is generally scaled to the same standard size.

Summary of the invention

The embodiments of the present application provide a method and device for processing pictures.

In a first aspect, an embodiment of the present application provides a method for processing a picture. The method includes: obtaining a picture to be processed; selecting a standard size corresponding to the picture to be processed from a preset standard size set as a target size; The size of the picture to be processed is adjusted to the target size to obtain the adjusted picture to be processed; the adjusted picture to be processed is input to a pre-trained picture processing model to obtain a processing result.

In some embodiments, selecting the standard size corresponding to the picture to be processed as the target size from the preset standard size set includes: determining the size of the picture to be processed; and from the preset standard size set based on the size of the picture to be processed Select the standard size corresponding to the picture to be processed as the target size.

In some embodiments, based on the size of the picture to be processed, selecting a standard size corresponding to the picture to be processed from the preset standard size set as the target size includes: selecting a standard size with the smallest difference from the size of the picture to be processed As the target size.

In some embodiments, the size of the picture to be processed is the resolution of the picture to be processed, and the resolution of the picture to be processed is expressed by the number of horizontal pixels and the number of vertical pixels; and for the standard size in the standard size set, the standard size and the The difference in the size of the processed picture is determined by the following steps: determining the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed; determining the ratio of the number of horizontal pixels to the number of vertical pixels corresponding to the standard size; The absolute value of the difference between the ratio of the number of pixels to the number of vertical pixels and the ratio of the number of horizontal pixels to the number of vertical pixels corresponding to the standard size is used as the difference between the standard size and the size of the picture to be processed.

In some embodiments, selecting a standard size corresponding to a picture to be processed as a target size from a preset standard size set includes: determining content information of the picture to be processed; and according to a pre-established correspondence between the content information and the standard size, A standard size corresponding to the content information with the highest similarity of the content information of the picture to be processed is determined as the target size.

In some embodiments, the image processing model is obtained by training as follows: obtaining a training sample set, wherein the training sample includes a sample picture and a processing result of the sample picture; and for the training samples in the training sample set, selecting a standard size set and The standard size corresponding to the sample picture in the training sample, and the size of the sample picture in the training sample is adjusted to the selected standard size to obtain the adjusted sample picture; using the method of machine learning, the The adjusted sample picture corresponding to the sample picture in the training sample is used as input, and the processing result corresponding to the input adjusted sample picture is used as the desired output, and a picture processing model is trained.

In a second aspect, an embodiment of the present application provides an apparatus for processing a picture. The apparatus includes: an obtaining unit configured to obtain a picture to be processed; and a selecting unit configured to select an image from a preset standard size set. The standard size corresponding to the picture to be processed is taken as the target size; the adjustment unit is configured to adjust the size of the picture to be processed to the target size to obtain the adjusted picture to be processed; the processing unit is configured to input the adjusted picture to be processed To the pre-trained image processing model, get the processing result.

In some embodiments, the above-mentioned selecting unit is further configured to: determine a size of the picture to be processed; and based on the size of the picture to be processed, select a standard size corresponding to the picture to be processed as a target size from a preset standard size set.

In some embodiments, the above-mentioned selecting unit is further configured to: select a standard size having a smallest difference from a size of a picture to be processed as a target size.

In some embodiments, the above-mentioned selecting unit is further configured to: determine content information of the picture to be processed; and determine content information with the highest similarity to the content information of the picture to be processed according to a pre-established correspondence between the content information and a standard size The corresponding standard size is used as the target size.

According to a third aspect, an embodiment of the present application provides an electronic device. The electronic device includes: one or more processors; a storage device configured to store one or more programs; The processor executes such that one or more processors implement the method as described in any implementation of the first aspect.

In a fourth aspect, an embodiment of the present application provides a computer-readable medium having stored thereon a computer program that, when executed by a processor, implements the method as described in any implementation manner of the first aspect.

The method and device for processing pictures provided by the embodiments of the present application obtain a picture to be processed and select a standard size corresponding to the picture to be processed from a preset standard size set as a target size. Then, the size of the picture to be processed is adjusted to the selected target size, and the adjusted picture is input to a pre-trained picture processing model to obtain a processing result, so that the size of the picture to be processed is adjusted by using preset multiple standard sizes. In this case, the size of the picture to be processed can be adjusted to the corresponding standard size according to the specific picture to be processed, which helps improve the flexibility of picture size adjustment.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features, objects, and advantages of the present application will become more apparent by reading the detailed description of the non-limiting embodiments with reference to the following drawings:

FIG. 1 is an exemplary system architecture diagram to which an embodiment of the present application can be applied; FIG.

2 is a flowchart of an embodiment of a method for processing a picture according to the present application;

3 is a flowchart of still another embodiment of a method for processing pictures according to the present application;

4 is a schematic diagram of an application scenario of a method for processing pictures according to the present application;

5 is a schematic structural diagram of an embodiment of an apparatus for processing pictures according to the present application;

FIG. 6 is a schematic structural diagram of a computer system suitable for implementing an electronic device according to an embodiment of the present application.

detailed description

The following describes the present application in detail with reference to the accompanying drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related invention, rather than limiting the invention. It should also be noted that, for convenience of description, only the parts related to the related invention are shown in the drawings.

It should be noted that, in the case of no conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The application will be described in detail below with reference to the drawings and embodiments.

FIG. 1 illustrates an exemplary architecture 100 of an embodiment of a method for processing pictures or an apparatus for processing pictures to which the present application can be applied.

As shown in FIG. 1, the system architecture 100 may include

terminal devices

101, 102, and 103, a network 104, and a server 105. The network 104 is a medium for providing a communication link between the

terminal devices

101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, among others.

The

terminal devices

101, 102, 103 interact with the server 105 through the network 104 to receive or send messages and the like. Various client applications can be installed on the

terminal devices

101, 102, and 103, such as a web browser application and an image processing application.

The

terminal devices

101, 102, and 103 may be hardware or software. When the

terminal devices

101, 102, and 103 are hardware, they can be various electronic devices capable of processing pictures, including but not limited to smart phones, tablet computers, e-book readers, laptop computers, and desktop computers. When the

terminal devices

101, 102, and 103 are software, they can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules (for example, multiple software or software modules used to provide distributed services), or it can be implemented as a single software or software module. It is not specifically limited here.

The server 105 may be a server that provides various services, for example, a picture processing server that provides analysis processing for pictures to be processed on the

terminal devices

101, 102, and 103. The picture processing server may perform processing such as resizing the picture to be processed, and generate a processing result, and may also feedback the processing result (such as the picture to be processed after the resizing) to the terminal device.

It should be noted that the above-mentioned pictures to be processed may also be directly stored locally on the server 105, and the server 105 may directly extract and process the locally stored pictures to be processed. At this time, the

terminal devices

101, 102, 103 and the network may not exist. 104.

It should be noted that the method for processing pictures provided by the embodiments of the present application is generally executed by the server 105, and accordingly, the apparatus for processing pictures is generally set in the server 105.

It should also be noted that the

terminal device

101, 102, 103 may also have an image processing application installed, and the

terminal device

101, 102, 103 may also process the image to be processed based on the image processing application. At this time, it is used to process the image The method can also be executed by the

terminal devices

101, 102, 103, and accordingly, the apparatus for processing pictures can also be set in the

terminal devices

101, 102, 103. At this time, the exemplary system architecture 100 may be absent from the server 105 and the network 104.

It should be noted that the server may be hardware or software. When the server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers or as a single server. When the server is software, it can be implemented as multiple software or software modules (for example, multiple software or software modules used to provide distributed services), or it can be implemented as a single software or software module. It is not specifically limited here.

It should be understood that the numbers of terminal devices, networks, and servers in FIG. 1 are merely exemplary. Depending on the implementation needs, there can be any number of terminal devices, networks, and servers.

With continued reference to FIG. 2, a flowchart 200 of an embodiment of a method for processing a picture according to the present application is shown. The method for processing pictures includes the following steps:

Step 201: Obtain a picture to be processed.

In this embodiment, the execution subject of the method for processing pictures (such as the server 105 shown in FIG. 1) may be wired or wirelessly from a local or other storage device (such as the

terminal devices

101 and 102 shown in FIG. 1). (103) Obtain a picture to be processed. The picture to be processed may be any type of picture.

In step 202, a standard size corresponding to a picture to be processed is selected as a target size from a preset standard size set.

In this embodiment, the size may refer to the physical size of the picture, such as the length and width of the picture. Size can also refer to the resolution of the picture. Size can also refer to the size of the picture, such as the storage space occupied by the picture. The standard size can be any preset size. According to different application requirements, different standard size sets can be preset.

The correspondence between the to-be-processed picture and the standard size may be different according to different correspondence logic. Different correspondences can usually use different selection methods to select the standard size from the preset standard size set as the target size.

In some optional implementations of this embodiment, the size of the picture to be processed may be determined first, and then based on the determined size of the picture to be processed, a standard size is selected as a target size from a preset standard size set.

Optionally, the size of the picture may be divided into several types in advance. For example, the size of a picture can be divided into three types: the aspect ratio is greater than 1, the aspect ratio is equal to 1, and the aspect ratio is less than 1. Then, for each type of size, a standard size corresponding to that type of size is determined. At this time, according to the size of the picture to be processed, a standard size corresponding to the size of the picture to be processed can be directly selected as a target size from a preset standard size set.

Specifically, for each type of size, a large number of pictures of that type can be obtained in advance, and then, for the obtained large number of pictures, the size of these pictures is adjusted to a standard size in a preset standard size set. After the average loss of accuracy. Similarly, an average value of the accuracy loss corresponding to each of the standard sizes in the preset standard size set after adjusting the sizes of these pictures is determined. Then, the standard size corresponding to the average value of the smallest loss of accuracy is selected as the standard size corresponding to the picture of this type.

In some optional implementations of this embodiment, the content information of the picture to be processed may be determined first, and then the similarity between the content information of the picture to be processed and the standard size will be similar to the content information of the picture to be processed. The standard size corresponding to the highest content information is the target size.

The content information may be information related to the content of the picture to be processed. For example, the content information may be the type of objects included in the picture to be processed, the number of objects included, and the like. Specifically, various existing picture detection or recognition methods may be used to determine content information of a picture to be processed. For example, existing deep learning models (such as AlexNet, etc.) used for picture classification can be used to classify pictures to be processed, and the tags of the pictures to be processed determined in the classification result are used as content information of the pictures to be processed. For example, some existing picture classification software can also be used, and the subject of the picture to be processed determined by the classification software can be used as the content information of the picture to be processed. For example, various existing picture understanding methods (such as image text description based on Multimodal Recurrent Neural Network) can be used to process the picture to be processed, and the obtained text describing the image can be used as content information of the picture to be processed.

For example, for a picture recognition application, a correspondence table that stores the correspondence between the content information of the picture and the standard size may be established in advance. For example, for a face picture, it may correspond to a standard size with an aspect ratio of 5: 3, and for a car picture, it may correspond to a standard size with an aspect ratio of 3: 1. At this time, when it is determined that the content information of the picture to be processed is a human face, a standard size of 5: 3 can be selected as the target size, and when the content information of the picture to be processed is a car, then a standard size of 3: 1 can be selected as the target size.

In some optional implementations of this embodiment, the standard size corresponding to the picture to be processed can be selected as the target size from the preset standard size set by the following steps:

In the first step, the size of the picture to be processed is adjusted to each standard size in the standard size set, so as to obtain the adjusted picture set to be processed.

In the second step, the similarity between each adjusted to-be-processed picture and the obtained to-be-processed picture in the adjusted to-be-processed picture set is calculated respectively, so as to obtain a similarity set.

In the third step, the maximum similarity is selected from the obtained similarity set, and the standard size corresponding to the selected maximum similarity is used as the target size.

Step 203: Adjust the size of the picture to be processed to the target size to obtain the adjusted picture to be processed.

In this embodiment, various applications for processing pictures can be used to adjust the size of the picture to be processed, thereby adjusting the size of the picture to be processed to the selected target size, and obtaining the resized picture to be processed.

Step 204: Input the adjusted to-be-processed picture into a pre-trained picture processing model to obtain a processing result.

In this embodiment, the picture processing model may be various models capable of processing pictures. For example, the picture processing model may be various deep learning models capable of processing pictures. The image processing model can also be various convolutional neural networks capable of processing images. The processing of pictures can be various processing methods. For example, picture classification, picture recognition, picture denoising, picture restoration, etc.

In some optional implementations of this embodiment, the picture processing model can be obtained by training as follows:

The first step is to obtain a training sample set, where the training sample includes a sample picture and a processing result of the sample picture.

In the second step, for the training samples in the training sample set, a standard size corresponding to the sample picture in the training sample is selected from the standard size set, and the size of the sample picture in the training sample is adjusted to the selected standard size. To get the adjusted sample picture.

Specifically, a standard size corresponding to a sample picture in each sample may be selected from a standard size set according to various methods for selecting a standard size described in step 202 in this embodiment.

The third step is to use machine learning to take the adjusted sample picture corresponding to the sample picture in the training sample in the training sample set as input, and use the processing result corresponding to the input adjusted sample picture as the desired output to train. The above-mentioned picture processing model capable of processing various standard sizes is obtained.

Specifically, the third step can be implemented by the following steps:

First, obtain the initial image processing model. Among them, the initial picture processing model may be various types of untrained or untrained artificial neural networks. For example, deep learning models. The initial picture processing model may also be a model obtained by combining a plurality of untrained or untrained artificial neural networks. For example, the initial picture processing model may be a model obtained by combining an untrained convolutional neural network, an untrained recurrent neural network, and an untrained fully connected layer. The initial image processing model can also be some existing, open source image processing models that have been trained and completed. The initial picture processing model may also be a neural network constructed by using various existing neural network APIs (Application Programming Interface).

Then, the adjusted sample picture corresponding to the sample picture in each training sample can be used as the input of the initial picture processing model, and the initial picture processing model is trained based on the output of the initial picture processing model and a preset loss function, and The initial image processing model after training is used as the image processing model.

The value of the loss function can be used to represent the degree of difference between the processing result output by the initial picture processing model and the processing result corresponding to the input adjusted sample picture during the training process. The smaller the loss function, the smaller the difference. And, it can be determined whether the training is completed according to the loss function. For example, the value of the loss function two or more times before and after may be less than a certain threshold as the end condition. At this time, when it is determined that the value of the loss function before and after two or more times is less than a certain threshold, the training is completed, and the initial image processing model after the training is completed is used as the trained image processing model.

In some optional implementations of this embodiment, a picture processing model corresponding to the standard size may be pre-trained for each standard size in a preset standard size set. At this time, after obtaining the adjusted picture to be processed, a pre-trained picture processing model corresponding to the target size may be selected according to the target size. The selected picture processing model is then used to process the adjusted picture to be processed. Specifically, for each standard size in the preset standard size set, a picture processing model corresponding to the standard size can be obtained by training as follows:

The first step is to obtain a training sample set, where the training samples may include sample pictures and processing results of the sample pictures.

In the second step, for each training sample in the training sample set, the size of the sample picture in the training sample is adjusted to the standard size to obtain the adjusted sample picture.

The third step is to use machine learning to take the adjusted sample picture corresponding to the sample picture in the training sample in the training sample set as input, and use the processing result corresponding to the input adjusted sample picture as the desired output to train. A picture processing model corresponding to the standard size is obtained.

The above third step can be specifically implemented by the following steps:

The value of the loss function can be used to represent the degree of difference between the processing result output by the initial picture processing model and the processing result corresponding to the input adjusted sample picture during the training process. The smaller the loss function, the smaller the difference. And, it can be determined whether the training is completed according to the loss function. For example, the value of the loss function two or more times before and after it may be less than a certain threshold value as the end condition. At this time, when it is determined that the value of the loss function before and after two or more times is less than a certain threshold value, the training is completed, and the initial picture processing model after the training is completed is used as the trained picture processing model.

In some optional implementations of this embodiment, the picture processing model can also be used to process pictures of other standard sizes except the target size in the standard size set, that is, the above picture processing model can process various preset standards. Size picture.

The method provided by the foregoing embodiment of the present application obtains a picture to be processed, and selects a standard size corresponding to the picture to be processed as a target size from a preset standard size set. Then, the size of the picture to be processed is adjusted to the selected target size, and the adjusted picture is input to a pre-trained picture processing model to obtain a processing result, so that the size of the picture to be processed is adjusted by using preset multiple standard sizes. In this case, the size of the picture to be processed can be adjusted to the corresponding standard size according to the specific picture to be processed, which helps improve the flexibility of picture size adjustment.

With further reference to FIG. 3, a flowchart 300 of yet another embodiment of a method for processing a picture is shown. The process 300 for processing pictures includes the following steps:

Step 301: Obtain a picture to be processed.

For the specific execution process of this step, refer to the related description of step 201 in the embodiment corresponding to FIG. 2, and details are not described herein again.

Step 302: Select a standard size that has the smallest difference from the size of the picture to be processed from the preset standard size set as the target size.

In this embodiment, the difference between each standard size in the preset standard size set and the size of the picture to be processed can be determined separately, and then the standard size corresponding to the smallest difference among them is selected as the target size. The difference between the size of the picture to be processed and the standard size can be determined using various calculation methods according to the actual application scenario.

Optionally, the difference between the length of the picture to be processed and the length corresponding to the standard size or the difference between the width of the picture to be processed and the width corresponding to the standard size may be directly used as the difference between the size of the picture to be processed and the standard size.

Optionally, the difference between the length of the picture to be processed and the length corresponding to the standard size may be calculated first, and may be recorded as the first difference; then, the difference between the width of the picture to be processed and the width corresponding to the standard size may be calculated, and Record it as the second difference; then, the square sum of the first difference and the second difference can be calculated, and the calculated sum of squares can be used as the difference between the size of the picture to be processed and the standard size.

Optionally, when the size of the picture is the resolution of the picture, the resolution may be expressed using the number of horizontal pixels and the number of vertical pixels. At this time, the difference between the sizes of the two pictures can be determined by the following steps.

In the first step, the size of the picture to be processed can be determined, and the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed can be calculated as the first ratio.

In the second step, the ratio of the number of horizontal pixels and the number of vertical pixels corresponding to the standard size to be compared with the size of the picture to be processed can be calculated as the second ratio.

In a third step, the absolute value of the difference between the first ratio and the second ratio is used as the difference between the sizes of the two pictures to be compared.

Step 303: Adjust the size of the picture to be processed to the target size to obtain the adjusted picture to be processed.

Step 304: Input the adjusted to-be-processed picture into a pre-trained picture processing model to obtain a processing result.

For specific implementation processes of the foregoing

steps

303 and 304, reference may be made to the related description of

steps

203 and 204 in the embodiment corresponding to FIG. 2, and details are not described herein again.

Further referring to FIG. 4, FIG. 4 is a schematic diagram of an application scenario of a method for processing a picture according to this embodiment. In the application scenario of FIG. 4, the resolution of the to-be-processed picture 401 is 320 * 480, where the number of horizontal pixels is 320 and the number of vertical pixels is 480. The preset standard size set 402 includes three standard sizes, which are: 240 * 320, 960 * 640, and 480 * 800. Then, the difference between the size of the picture to be processed and each standard size can be calculated separately.

Specifically, for a standard size of 240 * 320, a ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed may be calculated first as 320/480. Then, the ratio of the number of horizontal pixels to the number of vertical pixels of the standard size is calculated as 240/320. After that, the absolute value of the difference between the two ratios can be calculated as the difference between the size of the picture to be processed and the standard size 240 * 320 is 1/12, as shown by reference numeral 403 in the figure. Similarly, the difference between the size of the picture to be processed and the standard size 960 * 640 is 5/6, as shown by reference numeral 404 in the figure, and the difference between the size of the picture to be processed and the standard size 480 * 800 is 1 / 15, as shown by reference numeral 405 in the figure.

After that, the smallest difference value can be selected from the calculated three difference values to be 1/15. Therefore, the standard size 480 * 800 corresponding to the smallest difference is taken as the target size. Then, the size of the to-be-processed picture 401 can be adjusted to the above-mentioned target size 480 * 800 to obtain the adjusted to-be-processed picture 406.

Then, the adjusted to-be-processed picture 406 can be input to a pre-trained picture processing model 407, and a processing result 408 can be obtained. Alternatively, three picture processing models may be trained in advance, corresponding to three sizes in the standard size set 402, respectively. At this time, the adjusted to-be-processed picture is input to a picture processing model corresponding to the size of the adjusted to-be-processed picture to obtain a processing result.

As can be seen from FIG. 3, compared with the embodiment corresponding to FIG. 2, the flow 300 of the information pushing method in this embodiment can compare the difference between the size of the picture to be processed and each standard size in the standard size set, and The size of the picture to be processed is adjusted to the smallest difference corresponding to the standard size, and then the adjusted picture to be processed is processed to obtain a processing result. Therefore, compared with the solution described in this embodiment and the prior art in which the sizes of different pictures to be processed are uniformly adjusted to the same standard size, this method helps reduce the The degree of distortion or deformation further helps to reduce the subsequent impact on the processing result due to the distortion or deformation of the adjusted picture to be processed.

With further reference to FIG. 5, as an implementation of the methods shown in the foregoing figures, this application provides an embodiment of a device for processing pictures. This device embodiment corresponds to the method embodiment shown in FIG. 2, and the device is specific Can be applied to various electronic devices.

As shown in FIG. 5, the apparatus 500 for processing pictures provided in this embodiment includes an acquisition 501, a selection unit 502, an adjustment unit 503, and a processing unit 504. The obtaining unit 501 is configured to obtain a picture to be processed. The selecting unit 502 is configured to select a standard size corresponding to a picture to be processed as a target size from a preset standard size set. The adjustment unit 503 is configured to adjust the size of the picture to be processed to a target size to obtain the adjusted picture to be processed. The processing unit 504 is configured to input the adjusted to-be-processed picture to a pre-trained picture processing model to obtain a processing result.

In this embodiment, in the apparatus 500 for processing pictures: the specific processing of the obtaining unit 501, the selection unit 502, the adjustment unit 503, and the processing unit 504 and the technical effects brought by them can be referred to the corresponding embodiment in FIG. 2 respectively. The relevant descriptions of step 201, step 202, step 203, and step 204 are not repeated here.

In some optional implementations of this embodiment, the above-mentioned selecting unit 502 is further configured to: determine the size of the picture to be processed; and based on the size of the picture to be processed, select from the preset standard size set corresponding to the picture to be processed As the target size.

In some optional implementation manners of this embodiment, the above-mentioned selecting unit 502 is further configured to: select a standard size having a smallest difference from a size of a picture to be processed as a target size.

In some optional implementations of this embodiment, the size of the picture to be processed is the resolution of the picture to be processed, and the resolution of the picture to be processed is expressed by the number of horizontal pixels and the number of vertical pixels; and for the standards in the standard size set Size, the difference between the standard size and the size of the picture to be processed is determined by the following steps: determining the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed; determining the ratio of the number of horizontal pixels to the number of vertical pixels corresponding to the standard size; The absolute value of the difference between the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed and the ratio of the number of horizontal pixels and vertical pixels corresponding to the standard size is taken as the difference between the standard size and the size of the picture to be processed.

In some optional implementations of this embodiment, the above-mentioned selecting unit 502 is further configured to: determine content information of the picture to be processed; and determine the content of the picture to be processed according to the correspondence between the pre-established content information and the standard size The standard size corresponding to the content information with the highest similarity of information is taken as the target size.

In some optional implementations of this embodiment, the picture processing model is obtained by training as follows: obtaining a training sample set, where the training sample includes a sample picture and a processing result of the sample picture; and for the training samples in the training sample set, Select the standard size corresponding to the sample picture in the training sample from the standard size set, and adjust the size of the sample picture in the training sample to the selected standard size to obtain the adjusted sample picture; use the method of machine learning , Taking the adjusted sample picture corresponding to the sample picture in the training sample in the training sample set as input, and using the processing result corresponding to the input adjusted sample picture as the desired output, training to obtain a picture processing model.

The apparatus provided by the foregoing embodiments of the present application obtains a picture to be processed through an acquisition unit, the selection unit selects a standard size corresponding to the picture to be processed from a preset standard size set as a target size, and the adjustment unit adjusts the size of the picture to be processed To the target size, get the adjusted to-be-processed picture, and then the processing unit inputs the adjusted to-be-processed picture to a pre-trained picture processing model to get the processing result, so that the preset to-be-adjusted standard sizes are used to adjust the to-be-processed When the size of the picture is adjusted, the size of the picture to be processed can be adjusted to the corresponding standard size according to the specific picture to be processed, which helps to improve the flexibility of picture size adjustment.

Reference is now made to FIG. 6, which illustrates a schematic structural diagram of a computer system 600 suitable for implementing an electronic device according to an embodiment of the present application. The electronic device shown in FIG. 6 is only an example, and should not impose any limitation on the functions and scope of use of the embodiments of the present application.

As shown in FIG. 6, the computer system 600 includes a central processing unit (CPU) 601, which can be loaded into a random access memory (RAM) 603 from a program stored in a read-only memory (ROM) 602 or from a storage portion 608 Instead, perform various appropriate actions and processes. In the RAM 603, various programs and data required for the operation of the system 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

The following components are connected to the I / O interface 605: an input portion 606 including a keyboard, a mouse, and the like; an output portion 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage portion 608 including a hard disk and the like And a communication section 609 including a network interface card such as a LAN card, a modem, and the like. The communication section 609 performs communication processing via a network such as the Internet. The driver 610 is also connected to the I / O interface 605 as necessary. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage section 608 as needed.

In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart may be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product including a computer program carried on a computer-readable medium, the computer program containing program code for performing a method shown in a flowchart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 609, and / or installed from a removable medium 611. When the computer program is executed by a central processing unit (CPU) 601, the above-mentioned functions defined in the method of the present application are executed.

It should be noted that the computer-readable medium of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the foregoing. The computer-readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable Programming read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the foregoing. In this application, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal that is included in baseband or propagated as part of a carrier wave, and which carries computer-readable program code. Such a propagated data signal may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing. The computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, and the computer-readable medium may send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device . Program code embodied on a computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

The flowchart and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, which contains one or more functions to implement a specified logical function Executable instructions. It should also be noted that in some alternative implementations, the functions labeled in the blocks may also occur in a different order than those labeled in the drawings. For example, two blocks represented one after the other may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by a dedicated hardware-based system that performs the specified function or operation , Or it can be implemented with a combination of dedicated hardware and computer instructions.

The units described in the embodiments of the present application may be implemented by software or hardware. The described unit may also be provided in a processor, for example, it may be described as: a processor including an acquisition unit, a selection unit, an adjustment unit, and a processing unit. Among them, the names of these units do not constitute a limitation on the unit itself in some cases. For example, the obtaining unit may also be described as a “unit for obtaining a picture to be processed”.

As another aspect, the present application further provides a computer-readable medium, which may be included in the device described in the foregoing embodiments; or may exist alone without being assembled into the device. The computer readable medium carries one or more programs, and when the one or more programs are executed by the device, the device causes the device to: obtain a picture to be processed; and select a picture corresponding to the picture to be processed from a preset standard size set. The standard size is used as the target size; the size of the picture to be processed is adjusted to the target size to obtain the adjusted picture to be processed; the adjusted picture to be processed is input to a pre-trained picture processing model to obtain the processing result.

The above description is only a preferred embodiment of the present application and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solution of the specific combination of the above technical features, but should also cover the above technical features or Other technical solutions formed by arbitrarily combining their equivalent features. For example, a technical solution formed by replacing the above features with technical features disclosed in the present application (but not limited to) having similar functions.

Claims

A method for processing pictures, including:

Get the pending pictures;

Selecting a standard size corresponding to the picture to be processed as a target size from a preset standard size set;

Adjusting the size of the picture to be processed to the target size to obtain an adjusted picture to be processed;

The adjusted picture to be processed is input to a pre-trained picture processing model to obtain a processing result.
The method according to claim 1, wherein the selecting a standard size corresponding to the picture to be processed as a target size from a preset standard size set comprises:

Determining the size of the picture to be processed;

Based on the size of the picture to be processed, a standard size corresponding to the picture to be processed is selected as a target size from a preset standard size set.
The method according to claim 2, wherein, based on the size of the picture to be processed, selecting a standard size corresponding to the picture to be processed as a target size from a preset standard size set comprises:

A standard size with the smallest difference from the size of the picture to be processed is selected as the target size.
The method according to claim 3, wherein the size of the picture to be processed is the resolution of the picture to be processed, and the resolution of the picture to be processed is represented by the number of horizontal pixels and the number of vertical pixels; and

For a standard size in the standard size set, a difference between the standard size and the size of the picture to be processed is determined through the following steps:

Determining a ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed;

Determine the ratio of the number of horizontal pixels to the number of vertical pixels corresponding to the standard size;

Taking the absolute value of the difference between the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed and the ratio of the number of horizontal pixels and vertical pixels corresponding to the standard size as the difference between the standard size and the size of the picture to be processed value.
The method according to claim 1, wherein the selecting a standard size corresponding to the picture to be processed as a target size from a preset standard size set comprises:

Determining content information of the picture to be processed;

According to a pre-established correspondence between the content information and the standard size, a standard size corresponding to the content information with the highest similarity of the content information of the picture to be processed is determined as the target size.
The method according to any one of claims 1-5, wherein the picture processing model is obtained by training as follows:

Obtaining a training sample set, where the training sample includes a sample picture and a processing result of the sample picture;

For the training samples in the training sample set, select a standard size corresponding to the sample picture in the training sample from the standard size set, and adjust the size of the sample picture in the training sample to the selected standard size To get the adjusted sample picture;

By using a machine learning method, an adjusted sample picture corresponding to a sample picture in a training sample in the training sample set is taken as an input, and a processing result corresponding to the input adjusted sample picture is used as an expected output. The picture processing model is described.
A device for processing pictures includes:

An obtaining unit configured to obtain a picture to be processed;

A selection unit configured to select a standard size corresponding to the picture to be processed as a target size from a preset standard size set;

An adjustment unit configured to adjust a size of the picture to be processed to the target size to obtain an adjusted picture to be processed;

The processing unit is configured to input the adjusted to-be-processed picture into a pre-trained picture processing model to obtain a processing result.
The apparatus according to claim 7, wherein the selecting unit is further configured to:

Determining the size of the picture to be processed;

Based on the size of the picture to be processed, a standard size corresponding to the picture to be processed is selected as a target size from a preset standard size set.
The apparatus according to claim 8, wherein the selecting unit is further configured to select a standard size having a smallest difference from a size of the picture to be processed as a target size.
The apparatus according to claim 9, wherein the size of the picture to be processed is the resolution of the picture to be processed, and the resolution of the picture to be processed is expressed by the number of horizontal pixels and the number of vertical pixels; and

For a standard size in the standard size set, a difference between the standard size and the size of the picture to be processed is determined through the following steps:

Determining a ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed;

Determine the ratio of the number of horizontal pixels to the number of vertical pixels corresponding to the standard size;

Taking the absolute value of the difference between the ratio of the number of horizontal pixels to the number of vertical pixels of the picture to be processed and the ratio of the number of horizontal pixels and vertical pixels corresponding to the standard size as the difference between the standard size and the size of the picture to be processed value.
The apparatus according to claim 7, wherein the selecting unit is further configured to:

Determining content information of the picture to be processed;

According to a pre-established correspondence between the content information and the standard size, a standard size corresponding to the content information with the highest similarity of the content information of the picture to be processed is determined as the target size.
The apparatus according to any one of claims 7 to 11, wherein the picture processing model is obtained by training as follows:

Obtaining a training sample set, where the training sample includes a sample picture and a processing result of the sample picture;

For the training samples in the training sample set, select a standard size corresponding to the sample picture in the training sample from the standard size set, and adjust the size of the sample picture in the training sample to the selected standard size To get the adjusted sample picture;

By using a machine learning method, an adjusted sample picture corresponding to a sample picture in a training sample in the training sample set is taken as an input, and a processing result corresponding to the input adjusted sample picture is used as an expected output. The picture processing model is described.
An electronic device includes:

One or more processors;

A storage device on which one or more programs are stored;

When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1-6.
A computer-readable medium having stored thereon a computer program, wherein the program, when executed by a processor, implements the method of any one of claims 1-6.