WO2023239302A1 - 图像处理方法、装置、电子设备及存储介质 - Google Patents
图像处理方法、装置、电子设备及存储介质 Download PDFInfo
- Publication number
- WO2023239302A1 WO2023239302A1 PCT/SG2023/050401 SG2023050401W WO2023239302A1 WO 2023239302 A1 WO2023239302 A1 WO 2023239302A1 SG 2023050401 W SG2023050401 W SG 2023050401W WO 2023239302 A1 WO2023239302 A1 WO 2023239302A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- target
- feature
- encoding
- target attribute
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/094—Adversarial learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—Two-dimensional [2D] image generation
- G06T11/40—Filling planar surfaces by adding surface attributes, e.g. adding colours or textures
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—Two-dimensional [2D] image generation
- G06T11/60—Creating or editing images; Combining images with text
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Definitions
- the present disclosure provides an image processing method, device, electronic device and storage medium, which can change the overall attribute parameters of the image based on the characteristics of the image itself, making the effect of adding special effects more natural.
- Embodiments of the present disclosure provide an image processing method, which method includes: acquiring an image to be processed; inputting the image to be processed into an image attribute parameter change model to obtain a target image, wherein the target attribute parameter value of the target image is Different from the target attribute parameter value of the image to be processed, the target attribute parameter value of the target image and the target attribute parameter value of the image to be processed correspond to different target attribute expression states.
- the user can autonomously choose according to the prompt information whether to provide personal information to software or hardware such as electronic devices, applications, servers or storage media that perform the operations of the technical solution of the present disclosure.
- the method of sending prompt information to the user may be, for example, a pop-up window, and the prompt information may be presented in the form of text in the pop-up window.
- the pop-up window can also contain a selection control for the user to choose "agree” or "disagree” to provide personal information to the electronic device. It can be understood that the above process of notifying and obtaining user authorization is only illustrative and does not limit the implementation of the present disclosure.
- Figure 1 is a schematic flowchart of an image processing method provided by an embodiment of the present disclosure.
- the embodiment of the present disclosure is suitable for situations where image processing is performed to change image attributes based on the entire image to be processed.
- This method can be executed by an image processing device.
- the device can be implemented in the form of software and/or hardware, optionally, through electronic equipment, and the electronic equipment can be a mobile terminal, a personal computer (Personal Computer, PC) or a server, etc.
- the image processing method includes the following steps.
- the image to be processed can be an original image whose image attributes need to be changed, or an image obtained by downloading, shooting or uploading.
- the image to be processed can be obtained as the target image processing object.
- Image special effects that change the attributes of the image to be processed can be applied to any application that can process images or videos. It is understandable that during the video shooting process or the video call process, image special effects processing of changing the attributes of the image obtained in real time can also be performed.
- the method of this embodiment can make the combination effect of the change of the target attribute and the original image of the image to be processed more natural and vivid.
- the target attribute can be a pre-specified characteristic attribute of the object in the image to be processed, for example, the gender attribute, age attribute, hairstyle attribute, etc. of the person in the person image.
- the gender attribute as the target attribute as an example, assuming that the character object attribute parameters in the image to be processed are classified as male, the image to be processed can be input into the image attribute parameter change model, and the image attribute parameter change model outputs the target image, correspondingly , the target image is an image in which the gender attribute of the character object in the image to be processed is female.
- the output result of the image attribute parameter changing model in this embodiment can achieve the effect of "thousands of people with thousands of faces", because each picture has
- the processing process of the image to be processed is > ⁇ The hairstyle, hair color and facial features of the characters and objects in the image are personalized and generated.
- the diversity is greatly enriched and the degree of naturalness is also increased.
- the image attribute parameter change model may be a model trained based on image sample pairs including images with different target attribute expression states.
- the image of one target attribute expression state in the image sample pair can be used as the input of the trained model, and the image of another target attribute expression state can be used as the expected output image of the trained model, and the model training can be performed to obtain The image attribute parameter changing model applied in this embodiment.
- the image attribute parameter changing model can be obtained through model-training in a generative adversarial manner based on image sample pairs including images with different target attribute expression states.
- the Pix2Pix network architecture is used to train the model, so that the trained model can learn to "predict pixels based on pixels.”
- the image sample pair can be constructed by first generating multiple original sample images by a pre-trained image generator.
- the performance states of the target attributes in the multiple original sample images can be the same or different; then , perform feature encoding on each original sample image, and adjust the target attribute parameters (target attribute features) of the image feature encoding results to obtain the target encoding features; finally, the image generator that generates the original sample image decodes the target encoding features to obtain
- the target sample image corresponding to the original sample image is an image in which the target attribute performance state obtained after adjusting the target attribute characteristics is different from the original sample image.
- the construction process of the image sample pair is also based on the encoding process of the overall characteristics of the original sample image, and then the target attribute characteristics of the image feature encoding result are adjusted to finally obtain the corresponding image sample pair.
- the image attribute parameter change model can be Learn the inherent mapping relationship between pairs of image samples with different attribute representations.
- the pre-trained image generator can be an image generator obtained by training the StyleGAN model. into a device. During the training of StyleGAN, you can set the resolution of the generated image. For example, if you set the resolution of the generated image to 1024*1024, you can get multiple high-definition and high-quality original sample images.
- the model trained based on high-quality original sample images can also process high-quality images and obtain high-resolution image processing results.
- changing the target attributes in the model according to the image attribute parameters can make the output images of StyleGAN during the training process use images with different performance of the target attributes as a reference, and ultimately enable the StyleGAN model to be based on randomly sampled images.
- a vector of preset dimensions is generated to generate a collection of original sample images containing different performance states of the target attribute.
- the technical solution of the embodiment of the present disclosure is to obtain the target image by inputting the image to be processed into the image attribute parameter changing model after acquiring the image to be processed, wherein the target attribute parameter value of the target image is the same as the target attribute parameter value of the image to be processed.
- the target attribute parameter value of the target image and the target attribute parameter value of the image to be processed correspond to different target attribute performance states.
- the embodiment of the present disclosure uses the entire image to be processed as the processing object to change the image attributes, and solves the problem of using images in related technologies. It solves the problem of unnatural effect of changing image attributes by adding local special effects, and improves the effect of image processing, making the result of changing image attributes more natural.
- Figure 2 is a schematic flow chart of another image processing method provided by an embodiment of the present disclosure. In the process of implementing the method flow, the training process of the image attribute parameter changing model and the construction process of the training sample image are described.
- the method can be executed by an image processing device, which can be implemented in the form of software and/or hardware, optionally, by an electronic device, which can be a mobile terminal, a PC or a server. As shown in Figure 2, the image processing method includes the following steps.
- Step 1 Jointly train the image encoder based on the pre-trained image generator to obtain a target image encoding that enables the image feature encoding result of the original encoding object image to be decoded by the pre-trained image generator into the original encoding object image. device.
- the working process of the image attribute parameter changing model can be the encoding and decoding process of the image.
- a target image encoder is trained so that the image feature encoding result of the target image encoder can be pre-trained by the image decoder.
- StyleGAN model image generator, represented as Gs
- the image encoder First, input the original encoding object image (i.e., the original sample image) into the image encoder to obtain the image feature encoding vector of the original encoding object image; then, input the image feature encoding vector of the original encoding object image into the image generator and in the preset discriminator; finally, according to the feature decoded image generated by the image generator based on the image feature encoding vector of the original encoding object image and the preset discriminator compares the image feature encoding vector of the original encoding object image with the image generator
- the image encoder is updated with the discrimination result of the training sampling vector to obtain the target image encoder.
- the training process of the target image encoder when calculating the loss function during the training process of the image encoder, not only the first loss function between the feature decoded image and the original encoding image is calculated, but also a preset discriminator is set to determine the original image.
- the second loss function between the image feature encoding vector of the encoding object image and the training sampling vector of the image generator makes the image feature encoding vector of the original encoding object image close to the distribution characteristics of the training sampling vector of the image generator, improving the training effect.
- the training process of the image encoder is completed, and the target image encoder is obtained.
- the training process of the target image encoder can refer to the content shown in Figure 3. In Figure 3, two original sample images are shown.
- the two original sample images will be input to the encoder (Encoder) respectively.
- the encoder will perform feature encoding on the input image to obtain the feature W.
- a regularization processing mechanism is also added during the encoder-training process. N represents the number of regularization terms, such as - Represents the loss function in the regularization process.
- the encoder output result W is superposed with the regularization result, it is multiplied by the value N and then input into the pretrained StyleGAN for feature decoding to obtain the original encoding object image, and during the feature decoding process, input
- the special features in StyleGAN will also be input into the preset discriminator to determine the difference between the image feature encoding vector and the image generator's training sampling vector.
- the image feature encoding vector and the image generator's training sampling vector will suffer losses. Constraints on functions like.
- the training sampling vector is the image feature vector that the image generator randomly samples from the original sample image during the training process.
- Step 2 Use the target image encoder to encode each of the multiple original sample images with different target attribute performance states generated by the pre-trained image generator, and encode based on the image features of the multiple original sample images.
- the target attribute feature vector is determined.
- the feature vector corresponding to the target attribute can be determined.
- the original sample image is an image with a known label.
- the image feature encoding result (expressed as W) after the target image encoder encodes each original sample image also has a corresponding label.
- the image feature encoding result of each original sample image corresponds to the label W ⁇ male (male) or W ⁇ female (female).
- the feature vector representing the target attribute can be extracted from the classification learning of the image feature encoding results of multiple original sample images.
- SVM support vector machine classifier Machine
- W' target attribute feature vector
- the image to be processed can be directly input into the image attribute parameter change model to obtain the target image, where the target attribute parameter value of the target image is the same as the value of the image to be processed.
- the target attribute parameter values of are different.
- the target attribute parameter value of the target image and the target attribute parameter value of the image to be processed correspond to different target attribute performance states.
- the embodiment of the present application uses the entire image to be processed as the processing object to change the image attributes and solve the problem.
- FIG 4 is a schematic structural diagram of an image processing device provided by an embodiment of the present disclosure.
- the device is suitable for performing image processing to change image attributes based on the entire image to be processed. It can be implemented in the form of software and/or hardware.
- the image processing device can be configured in an electronic device, and the electronic device can be a mobile terminal, a PC, a server, etc. As shown in Figure 4, the image processing device includes: an image acquisition module 310 and an image processing module 320.
- Image acquisition module 310 configured to acquire the image to be processed;
- Image processing module 32. is set to input the image to be processed into the image attribute parameter changing model to obtain the target image, wherein the target attribute parameter value of the target image is different from the target attribute parameter value of the image to be processed, and the target attribute parameter value of the target image is
- the target attribute parameter value and the target attribute parameter value of the image to be processed correspond to different target attribute performance states.
- the technical solution provided by the embodiment of the present application obtains an image to be processed and inputs the image to be processed into an image attribute parameter changing model to obtain a target image, wherein the target attribute parameter value of the target image is the same as the target attribute parameter value to be processed.
- the target attribute parameter value of the image is different.
- the image sample pair includes an original sample image generated by a pre-trained image generator, and an image corresponding to the original sample image obtained by decoding the target encoding feature by the pre-trained image generator.
- Target sample image wherein the target encoding feature is a feature obtained by performing feature encoding on the original sample image and adjusting the target attribute parameters of the image feature encoding result.
- the image processing device further includes a training sample construction module, configured to: jointly train the image encoder based on the pre-trained image generator to obtain an image that can make the original encoding object
- the image feature encoding result is decoded by the pre-trained image generator into the target image encoder of the original encoding object image; the target attribute generated by the pre-trained image generator is processed by the target image encoder Perform feature encoding on each of the multiple original sample images with different performance states, and determine the target attribute feature vector based on the image feature encoding results of the multiple original sample images;
- Based on the target attribute feature vector pair Edit the target attribute parameters of the image feature encoding results of each original sample image to obtain the changed image feature encoding results of each original sample image; input the changed image feature encoding results of each original sample image into the In the above-mentioned pre-trained image generator, a target image corresponding to each of the original sample images is obtained, and an image sample pair is generated.
- the training sample construction module is configured to: input the original encoding object image into the image encoder to obtain the image feature encoding vector of the original encoding object image;
- the image feature encoding vector of the original encoding object image is input into the pre-trained image generator and the preset discriminator respectively; according to the pre-trained image generator, the image feature encoding vector of the original encoding object image is generated.
- the feature decoded image and the preset discriminator update the image encoder with the discrimination result between the image feature encoding vector and the training sampling vector of the pre-trained image generator to obtain a target image encoder.
- the training sample construction module is configured to: classify the image feature encoding results of the multiple original sample images through a support vector machine classifier; determine based on the classification results to make the The target attributes of multiple original sample images represent target attribute feature vectors with different states.
- the training sample construction module is configured to: determine the target attribute feature vector according to the target attribute performance state of the original sample image corresponding to the image feature encoding result of each original sample image. Corresponding attribute editing weight value; change the value of each original sample image The image feature encoding result is superimposed on the product of the target attribute feature vector and the attribute editing weight value to obtain the modified image feature encoding result of each original sample image.
- FIG. 5 is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. Referring now to FIG. 5 , an electronic device (such as the terminal device or server in FIG. 5 ) 40 suitable for implementing embodiments of the present disclosure is shown. Structural diagram.
- Embodiments of the present disclosure provide a computer storage medium on which a computer program is stored. When the program is executed by a processor, the image processing method provided by the above embodiments is implemented.
- the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the above two.
- the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above.
- Examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard drives, RAM. ROM. Erasable Programmable Read-Only Memory , EPROM) or flash memory, optical fiber, portable compact disk read-only memory (Compact Disc Read-Only Memory, CD-ROM).
- a computer-readable storage medium may be any tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device.
- the computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, in which computer-readable program code is carried. This propagated data signal can take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above.
- a computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device .
- the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, radio frequency (Radio Frequency, RF), etc., or any suitable combination of the above.
- the client and server can communicate using any currently known or future developed network protocol such as HyperText Transfer Protocol (HTTP), and can communicate with any form or medium.
- Digital data communications e.g., communications networks interconnections. Examples of communication networks include Local Area Networks (LANs), Wide Area Networks (WANs), the Internet (eg, the Internet), and end-to-end networks (eg, ad hoc end-to-end networks), as well as any current network for knowledge or future research and development.
- Computer program code for performing operations of the present disclosure may be written in one or more programming languages, including but not limited to object-oriented programming languages such as Java, Smalltalk > C++, or a combination thereof , also includes conventional procedural programming languages such as "Such as” or similar programming languages.
- the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly Execute partially on the user's computer on a remote computer, or execute entirely on the remote computer or server.
- the remote computer can be connected to the user's computer through any kind of network—including a LAN or WAN—or , can be connected to an external computer (such as through the Internet using an Internet service provider).
- each box in a flowchart or block diagram may represent a module, segment, or portion of code.
- the module, segment, or portion of code Contains one or more executable instructions for implementing specified logical functions.
- the functions noted in the block may occur in a sequence different from that noted in the figures. For example, two blocks shown one after another may actually execute substantially in parallel, or they may execute in the reverse order, depending upon the functionality involved.
- a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device.
- the machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
- Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any appropriate combination of the foregoing.
- Example 1 provides an image processing method, which method includes: acquiring an image to be processed; inputting the image to be processed into an image attribute parameter change model to obtain a target image, wherein the target attribute parameter value of the target image is different from the target attribute parameter value of the image to be processed, and the target attribute parameter value of the target image and the target attribute parameter value of the image to be processed correspond to different targets. Properties represent status.
- the pre-trained image generator decodes the target coding feature to obtain the target sample image corresponding to the original sample image, wherein the target coding feature is the result of feature encoding the original sample image and encoding the image feature Features obtained by adjusting target attribute parameters; wherein, the target attribute expression state of the original sample image is different from the target attribute expression state of the target sample image corresponding to the original sample image.
- Example 3 provides an image processing method, including:
- the number of original sample images is multiple, and the multiple original sample images are the pre-trained
- the image generator generates multiple original sample images with different target attribute performance states;
- the generation process of the image sample pair includes: jointly scraping the image encoder based on the pre-scratched image generator to obtain an image that can be used
- the image feature encoding result of the original encoding object image is decoded by the pre-trained image generator into the target image encoder of the original encoding object image; and the target image encoder generates the image feature encoding result of the pre-trained image generator.
- Each original sample image in the plurality of original images with different target attribute expression states is feature encoded, and a target attribute feature vector is determined based on the image feature encoding results of the multiple original sample images; based on the target attribute feature vector Edit the target attribute parameters of the image feature encoding results of each original sample image to obtain the changed image feature encoding results of each original sample image; input the changed image feature encoding results of each original sample image into In the pre-trained image generator, a target image corresponding to each original sample image is obtained, and the image sample pair is generated.
- Example 4 provides an image processing method, further comprising:
- encoding the image based on the pre-trained image generator perform joint training to obtain a target image encoder that enables the image feature encoding result of the original encoding object image to be decoded by the pre-trained image generator into the original encoding object image, including: converting the original encoding object image The object image is input into the image encoder to obtain the image feature encoding vector of the original encoded object image >; the image feature encoding vector of the original encoded object image is input to the pre-trained image generator and preset discrimination respectively.
- the device according to the feature decoding image generated by the pre-trained image generator based on the image feature encoding vector of the original encoding object image and the preset discriminator comparing the image feature encoding vector of the original encoding object image with the The image encoder is updated with the discrimination result of the training sampling vector of the pre-trained image generator to obtain the target image encoder.
- Example 6 provides an image processing method, further comprising:
- the feature vector based on the target attribute Edit the target attribute parameters on the image feature encoding result of each original sample image to obtain the changed image feature encoding result of each original sample image, including: the original image feature encoding result corresponding to each original sample image
- the target attribute performance state of the sample image determines the attribute editing weight value corresponding to the target attribute feature vector; superimposes the image feature encoding result of each original sample image with the product of the target attribute feature vector and the attribute editing weight value , to obtain the modified image feature encoding result of each original sample image.
- the image processing device also includes:
- the number of the original sample images is multiple, and the multiple original sample images are target attributes with different performance states generated by the pre-trained image generator.
- a plurality of original sample images; the image processing device also includes a training sample construction module, configured to: jointly train the image encoder based on the pre-trained image generator to obtain image features that enable the original encoding object image The encoding result is decoded by the pre-trained image generator into a target image encoder of the original encoding target image; and the target attribute representation state generated by the pre-trained image generator is generated by the target image encoder.
- each original sample image is Edit the target attribute parameters of the image feature coding results to obtain the changed image feature coding results of each original sample image; input the changed image feature coding results of each original sample image into the pre-trained image In the generator, a target image corresponding to each original sample image is obtained, and the image sample pair is generated.
- Example 10 provides an image processing device, further comprising:
- the training sample construction module is configured to: The original encoding object image is input into the image encoder to obtain the image feature encoding vector of the original encoding object image; the image feature encoding vector of the original encoding object image is input to the pre-trained image generator and preset respectively.
- Example 12J provides an image processing device, further comprising:
- the training sample construction module is configured to: according to The image feature encoding result of each original sample image corresponds to the target attribute performance status of the original sample image. Statefully determine the attribute editing weight value corresponding to the target attribute feature vector; superimpose the image feature encoding result of each original sample image with the product of the target attribute feature vector and the attribute editing weight value to obtain each of the The modified image feature encoding result of the original sample image.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
Description
Claims
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US18/873,469 US20260030510A1 (en) | 2022-06-10 | 2023-06-05 | Image processing method and apparatus, electronic device and storage medium |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210658013.2A CN117252954A (zh) | 2022-06-10 | 2022-06-10 | 图像处理方法、装置、电子设备及存储介质 |
| CN202210658013.2 | 2022-06-10 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023239302A1 true WO2023239302A1 (zh) | 2023-12-14 |
Family
ID=89118659
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/SG2023/050401 Ceased WO2023239302A1 (zh) | 2022-06-10 | 2023-06-05 | 图像处理方法、装置、电子设备及存储介质 |
Country Status (3)
| Country | Link |
|---|---|
| US (1) | US20260030510A1 (zh) |
| CN (1) | CN117252954A (zh) |
| WO (1) | WO2023239302A1 (zh) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111292262A (zh) * | 2020-01-19 | 2020-06-16 | 腾讯科技(深圳)有限公司 | 图像处理方法、装置、电子设备以及存储介质 |
| CN111325657A (zh) * | 2020-02-18 | 2020-06-23 | 北京奇艺世纪科技有限公司 | 图像处理方法、装置、电子设备和计算机可读存储介质 |
| CN112241741A (zh) * | 2020-08-25 | 2021-01-19 | 华中农业大学 | 基于分类对抗网的自适应图像属性编辑模型和编辑方法 |
| CN114239717A (zh) * | 2021-12-15 | 2022-03-25 | 北京欧珀通信有限公司 | 模型训练方法、图像处理方法及装置、电子设备、介质 |
-
2022
- 2022-06-10 CN CN202210658013.2A patent/CN117252954A/zh active Pending
-
2023
- 2023-06-05 US US18/873,469 patent/US20260030510A1/en active Pending
- 2023-06-05 WO PCT/SG2023/050401 patent/WO2023239302A1/zh not_active Ceased
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111292262A (zh) * | 2020-01-19 | 2020-06-16 | 腾讯科技(深圳)有限公司 | 图像处理方法、装置、电子设备以及存储介质 |
| CN111325657A (zh) * | 2020-02-18 | 2020-06-23 | 北京奇艺世纪科技有限公司 | 图像处理方法、装置、电子设备和计算机可读存储介质 |
| CN112241741A (zh) * | 2020-08-25 | 2021-01-19 | 华中农业大学 | 基于分类对抗网的自适应图像属性编辑模型和编辑方法 |
| CN114239717A (zh) * | 2021-12-15 | 2022-03-25 | 北京欧珀通信有限公司 | 模型训练方法、图像处理方法及装置、电子设备、介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| US20260030510A1 (en) | 2026-01-29 |
| CN117252954A (zh) | 2023-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20260030809A1 (en) | Image processing method and apparatus, storage medium, and electronic device | |
| WO2023125361A1 (zh) | 文字生成方法、装置、电子设备及存储介质 | |
| WO2023138560A1 (zh) | 风格化图像生成方法、装置、电子设备及存储介质 | |
| CN112581635B (zh) | 一种通用的快速换脸方法、装置、电子设备和存储介质 | |
| CN114049417B (zh) | 虚拟角色图像的生成方法、装置、可读介质及电子设备 | |
| CN111967397A (zh) | 人脸影像处理方法和装置、存储介质和电子设备 | |
| US20250095112A1 (en) | Image processing method and apparatus, device, and medium | |
| WO2023202543A1 (zh) | 文字处理方法、装置、电子设备及存储介质 | |
| US20240386640A1 (en) | Method, apparatus, device and storage medium for generating character style profile image | |
| CN114863214A (zh) | 图像生成模型训练、图像生成方法、装置、介质及设备 | |
| CN116596748A (zh) | 图像风格化处理方法、装置、设备、存储介质和程序产品 | |
| WO2024037556A1 (zh) | 图像处理方法、装置、设备及存储介质 | |
| CN115546575A (zh) | 驱动模型的训练方法、驱动方法、装置、可读介质和设备 | |
| WO2023231918A1 (zh) | 图像处理方法、装置、电子设备及存储介质 | |
| WO2023140787A2 (zh) | 视频的处理方法、装置、电子设备、存储介质和程序产品 | |
| CN111402133A (zh) | 图像处理方法、装置、电子设备及计算机可读介质 | |
| CN111402113A (zh) | 图像处理方法、装置、电子设备及计算机可读介质 | |
| WO2023027628A1 (zh) | 图像生成方法、装置、设备及存储介质 | |
| CN111933122B (zh) | 语音识别方法、装置、电子设备和计算机可读介质 | |
| WO2023239302A1 (zh) | 图像处理方法、装置、电子设备及存储介质 | |
| EP4435724A2 (en) | Video generation method and device | |
| US20250356578A1 (en) | Image processing method and apparatus, electronic device, and storage medium | |
| CN115937338B (zh) | 图像处理方法、装置、设备及介质 | |
| CN116030321A (zh) | 生成图像的方法、装置、电子设备和存储介质 | |
| CN115953597A (zh) | 图像处理方法、装置、设备及介质 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23820197 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 18873469 Country of ref document: US |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 23820197 Country of ref document: EP Kind code of ref document: A1 |
|
| WWP | Wipo information: published in national office |
Ref document number: 18873469 Country of ref document: US |
