WO2023010753A1 - 一种色域映射方法、装置、终端设备及存储介质 - Google Patents

一种色域映射方法、装置、终端设备及存储介质 Download PDF

Info

Publication number
WO2023010753A1
WO2023010753A1 PCT/CN2021/138136 CN2021138136W WO2023010753A1 WO 2023010753 A1 WO2023010753 A1 WO 2023010753A1 CN 2021138136 W CN2021138136 W CN 2021138136W WO 2023010753 A1 WO2023010753 A1 WO 2023010753A1
Authority
WO
WIPO (PCT)
Prior art keywords
color gamut
color
video
map
full convolution
Prior art date
Application number
PCT/CN2021/138136
Other languages
English (en)
French (fr)
Inventor
章政文
陈翔宇
董超
乔宇
Original Assignee
中国科学院深圳先进技术研究院
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 中国科学院深圳先进技术研究院 filed Critical 中国科学院深圳先进技术研究院
Publication of WO2023010753A1 publication Critical patent/WO2023010753A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/92—Dynamic range modification of images or parts thereof based on global image properties
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H04N21/2343—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements
    • H04N21/234309—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements by transcoding between formats or standards, e.g. from MPEG-2 to MPEG-4 or from Quicktime to Realvideo
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H04N21/4402—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display
    • H04N21/440218—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving reformatting operations of video signals for household redistribution, storage or real-time display by transcoding between formats or standards, e.g. from MPEG-2 to MPEG-4
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N7/00—Television systems
    • H04N7/01—Conversion of standards, e.g. involving analogue television standards or digital television standards processed at pixel level
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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; CALCULATING OR 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 application relates to the technical field of image processing, and in particular to a color gamut mapping method, device, terminal equipment and storage medium.
  • the same original image or original video will have different effects (such as brightness, color, contrast, saturation, etc.) under different display devices, because the color areas used by different display devices (That is, the color gamut) is different, and the source color gamut of the image is often converted to the target color gamut by means of color gamut mapping, so that the effect of the same original image or original video displayed on different display devices is as consistent as possible.
  • transfer functions are often used to implement gamut mapping. That is, by measuring the calibration value between the two color gamuts, and determining a conversion function based on the calibration value, and then using the conversion function to convert the color value of each color node in the source color gamut to the target color gamut. Since the color distribution of different color gamut color spaces is not evenly distributed, using the same conversion function for each color node will produce a certain degree of color shift, resulting in poor color accuracy of the converted image.
  • the present application provides a color gamut mapping method, device, terminal equipment, and storage medium, which can improve the color accuracy of image conversion during the color gamut mapping process.
  • the present application provides a color gamut mapping method, the method comprising: acquiring a first color gamut map, the first color gamut map including a plurality of color node values of the first color gamut;
  • the first color gamut map is input into the trained full convolution model for processing, and the second color gamut map is output, and the second color gamut map includes color node values corresponding to the plurality of color node values one-to-one
  • the fully convolutional model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1
  • N-1 activation functions are interspersed in the N convolutional layers, and N is an integer greater than or equal to 3.
  • the activation function is a non-linear activation function.
  • the training method of the full convolution model includes: using a preset training set and a preset loss function to iteratively train the full convolution initial model to obtain the full convolution model;
  • the training set includes a plurality of first color gamut map samples and a second color gamut map sample corresponding to each of the first color gamut map samples;
  • the plurality of first color gamut map samples is at least one first color gamut sample a video frame in a video sample, the second color gamut map sample being a video frame in a second color gamut video sample corresponding to the first color gamut video sample;
  • the preset loss function is used to describe the L2 loss between the predicted second color gamut map and the sample of the second color gamut map, and the predicted second color gamut map is the pair of the full convolution model obtained by processing the first color gamut map sample.
  • the method further includes: determining a color lookup table between the first color gamut map and the second color gamut map according to the first color gamut map and the second color gamut map.
  • the first color gamut is the SDR color gamut
  • the second color gamut is the HDR color gamut
  • the application method of the color lookup table includes:
  • the present application provides a color gamut mapping device, including:
  • An acquisition unit configured to acquire a first color gamut map, the first color gamut map including a plurality of color node values of the first color gamut;
  • a processing unit configured to input the first color gamut map into the trained full convolution model for processing, and output a second color gamut map, the second color gamut map includes values corresponding to the plurality of color nodes one by one
  • the full convolution model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1, and N-1 activation functions are interspersed in the N convolutional layers, and N is greater than or An integer equal to 3.
  • the present application provides a terminal device, including: a memory and a processor, where the memory is used to store a computer program; and the processor is used to execute the method described in any one of the above first aspects when calling the computer program.
  • the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in any one of the above-mentioned first aspects is implemented.
  • an embodiment of the present application provides a computer program product, which, when the computer program product runs on a processor, causes the processor to execute the method described in any one of the above-mentioned first aspects.
  • a color gamut mapping method, device, terminal equipment, and storage medium provided in this application use a full convolution model to implement color gamut mapping.
  • the full convolution model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1 and N-1 activation functions interspersed, so that the values of multiple color nodes in the first color gamut map have their corresponding color values. Domain mapping function. Therefore, the second color gamut map processed by the full convolution model includes color node values corresponding to the multiple color node values in the first color gamut map one-to-one, avoiding the problem caused by multiple color node values in the first color gamut map All color node values use the same conversion function to cause color shift. Therefore, using the color gamut mapping method provided by this application can realize non-uniform color gamut mapping and improve the color accuracy of image conversion during the color gamut mapping process.
  • FIG. 1 is a schematic flowchart of a color gamut mapping method provided by an embodiment of the present application
  • Fig. 2 is an architecture diagram of a full convolution model of a color gamut mapping method provided by an embodiment of the present application
  • Figure 3 is a schematic diagram of the representation ranges of the HDR color gamut and the SDR color gamut provided by an embodiment of the present application
  • FIG. 4 is a schematic diagram of a color gamut mapping device provided by an embodiment of the present application.
  • Fig. 5 is a schematic structural diagram of a terminal device provided by an embodiment of the present application.
  • gamut mapping is usually implemented using calibration methods. That is, by measuring the calibration value between the two color gamuts, and determining a conversion function based on the calibration value, and then using the conversion function to convert the color value of each color node in the source color gamut to the target color gamut. Since the color distribution of different color gamut color spaces is not evenly distributed, using the same conversion function for each color node will produce a certain degree of color shift, resulting in poor color accuracy of the converted image.
  • the present application provides a color gamut mapping method, which uses a full convolution model to implement color gamut mapping.
  • the full convolution model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1 and N-1 activation functions interspersed, so that the values of multiple color nodes in the first color gamut map have their corresponding color values. Domain mapping function. Therefore, the second color gamut map processed by the full convolution model includes color node values corresponding to the multiple color node values in the first color gamut map one-to-one, avoiding the problem caused by multiple color node values in the first color gamut map All color node values use the same conversion function to cause color shift. Therefore, using the color gamut mapping method provided by this application can realize non-uniform color gamut mapping and improve the color accuracy of image conversion during the color gamut mapping process.
  • FIG. 1 is a flowchart of an embodiment of a color gamut mapping method provided by the present application.
  • the executor of the color gamut mapping method provided in the embodiment of the present application may be an image/video processing device, wherein the image/video processing device may be a mobile terminal device such as a smart phone, a tablet computer, a camera, or a desktop computer, a robot, Servers and other terminal devices capable of processing image/video data.
  • the image/video processing device may be a mobile terminal device such as a smart phone, a tablet computer, a camera, or a desktop computer, a robot, Servers and other terminal devices capable of processing image/video data.
  • the color gamut mapping method in the embodiment of the present application includes:
  • the first color gamut may be BT.709 color gamut, BT.2020 color gamut, or DCI-P3 color gamut, etc.
  • the first color gamut diagram may be composed of multiple color node values in the first color gamut image.
  • the first color gamut map may be an image of a color node value in the first color gamut captured, downloaded, or read from a local storage area, or may be intercepted from a video whose color node value is in the first color gamut Video frame or image.
  • the first color gamut map can also be a synthesized image, for example, obtain all the color node values in the BT.709 color gamut, and use an image synthesis tool to synthesize all the color node values in the BT.709 color gamut obtained above image.
  • the second color gamut map includes color nodes corresponding to a plurality of color node values of the first color gamut one-to-one value
  • the fully convolutional model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1
  • N-1 activation functions are interspersed in the N convolutional layers
  • N is an integer greater than or equal to 3.
  • N is an integer greater than or equal to 3.
  • the activation function may be a linear rectification function (Rectified Linear Unit, ReLU). Choosing the ReLU activation function as the activation function in the full convolution model can not only speed up the calculation efficiency, but also increase the nonlinear fitting ability of the full convolution model. Certainly, the activation function may also be other types of activation functions, for example, a Sigmoid function or a Thnh function.
  • this embodiment adopts a full convolution model including 3 convolution layers with a convolution kernel size of 1 ⁇ 1, and 2 ReLU activation functions interspersed among the 3 convolution layers.
  • the first color gamut image is input into the full convolution model shown in Figure 2 for processing, and the second color gamut image can be output.
  • the full convolution model provided by this application is composed of N convolution layers with a convolution kernel size of 1 ⁇ 1 and N-1 activation functions interspersed, so that multiple Each color node value has its own corresponding color gamut mapping function, so that the second color gamut map processed by the full convolution model includes color node values corresponding to multiple color node values in the first color gamut map one-to-one , to avoid the color shift due to the use of the same conversion function for multiple color node values in the first color gamut map, therefore, the color gamut mapping method provided by this application can improve the accuracy of image conversion in the color gamut mapping process color standard.
  • the full convolutional model is composed of N convolutional layers with a convolution kernel size of 1 ⁇ 1 and N-1 activation functions interspersed, the model structure is simple and the number of parameters used is relatively small, which can effectively reduce the The computational cost of the fully convolutional model processing tasks improves computational efficiency and speeds up task processing.
  • a color lookup table (color lookup table) can also be made by using the full convolution model provided in this application. That is, according to the first color gamut map and the second color gamut map generated by using the full convolution model, a color lookup table between the first color gamut map and the second color gamut map is determined.
  • the color lookup table can be directly added to the post-processing process of terminal devices such as cameras to improve the quality of images or videos captured by terminal devices such as cameras from the perspective of software.
  • the color lookup table can also be applied in the image/video editor as a means of image or video post-processing to improve the color accuracy of the image or video conversion during the color gamut mapping process, for example, using the color lookup table to achieve different filter images mirror effect.
  • the color lookup table can also be used for color gamut mapping between different display devices, so that the same image/video can display the same effect as possible on different display devices.
  • the preset training set and the preset loss function are used to iteratively train the full convolution initial model to obtain the full convolution model.
  • the training set includes a plurality of first color gamut map samples and a second color gamut map sample corresponding to each first color gamut map sample.
  • the acquisition method (or source) of the first color gamut map sample can be the image data whose color node value is in the first color gamut directly acquired through video or image acquisition equipment, or the color to be acquired Image data converted from video data with node values in the first color gamut by frame extraction or frame splitting.
  • the second color gamut map sample can also be the image data whose color node values are in the second color gamut directly obtained through video or image acquisition equipment, or it can also be the video data with the acquired color node values in the second color gamut for frame extraction Or the image data converted by frame splitting and other methods.
  • the plurality of first color gamut map samples are video frames in at least one first color gamut video sample
  • the second color gamut map samples are video frames in a second color gamut video sample corresponding to the first color gamut video sample. video frames.
  • the training set uses the first color gamut video samples and the corresponding second color gamut video samples to extract more abundant Color node values, using video frames with richer color node values (including image samples of the first color gamut and image samples of the second color gamut) to train the full convolution model can also improve the accuracy of model training.
  • the preset loss function is used to describe the L2 loss between the predicted second color gamut map and the second color gamut map sample, and the predicted second color gamut map is obtained by processing the first color gamut map sample with the full convolution model of.
  • the initial model can be trained by designing the corresponding training set and loss function, so as to obtain a fully convolutional model suitable for different color gamut mapping tasks.
  • the training process and application of the full convolution model provided by the present application will be exemplarily described below.
  • FIG. 3 it is a schematic diagram of the representation range of HDR color gamut and SDR color gamut.
  • BT.709 and BT.2020 are both TV parameter standards issued by ITU (International Telecommunication Union)
  • DCI-P3 is the digital The color gamut standard developed by movie theaters.
  • BT.2020 has the largest range among DCI-P3, BT.709 and BT.2020
  • the color gamut range of DCI-P3 is second
  • the color gamut range represented by BT.709 is the smallest.
  • SDR images/videos use the BT.709 color gamut
  • HDR images/videos use the wider BT.2020 color gamut or DCI-P3 color gamut.
  • the HDR image/video can show higher contrast and richer colors than the SDR image/video.
  • a training set is acquired, and the training set may include multiple SDR video frame samples and HDR video frame samples corresponding to the multiple SDR video frame samples one-to-one.
  • an SDR video sample and its corresponding HDR video sample are acquired first.
  • SDR video samples and corresponding HDR video samples can be obtained from public video websites. It is also possible to perform SDR and HDR processing on videos in the same RAW data format, respectively, to obtain SDR video samples and corresponding HDR video samples. It is also possible to use the SDR camera and the HDR camera respectively to shoot corresponding SDR video samples and HDR video samples in the same scene.
  • the SDR video samples and their corresponding HDR video samples are frame-drawn to obtain a plurality of SDR video frame samples and the temporal and spatial connections between multiple SDR video samples.
  • a frame extraction tool can be used to extract frames from the SDR video sample and its corresponding HDR video sample.
  • FFmpeg Fast Forward Mpeg
  • the fully convolutional initial model is iteratively trained using the preset training set and the preset loss function to obtain a fully convolutional model.
  • the preset loss function is used for L2 loss between multiple HDR video frames predicted by the fully convolutional inception model and HDR video frame samples.
  • the full convolution initial model can be iteratively trained by the gradient descent method until the model converges, and the trained full convolution model can be obtained.
  • the color mapping table can be obtained based on the full convolution model.
  • the SDR color gamut map is generated according to all the color node values in the SDR color gamut, and after being input to the full convolution model for processing, the corresponding HDR color gamut map can be obtained. Then establish a corresponding relationship between the color nodes in the SDR color gamut map and the HDR color gamut map to obtain a color lookup table.
  • an embodiment of the present application provides an HDR video conversion method, the method comprising:
  • the acquisition method of the SDR video to be processed can be a complete video taken, downloaded or read from a local storage area, or an SDR video segment intercepted from a completed video, or it can be obtained by using different color gamuts. All or part of the color node values synthesize different color gamut maps.
  • the SDR video to be processed adopts an 8-bit encoding format
  • the HDR color gamut video obtained after color mapping by the color lookup table also adopts an 8-bit encoding format.
  • the encoding format of the HDR color gamut is converted to the encoding format of the HDR video to obtain the HDR video corresponding to the SDR video.
  • the 8-bit encoded HDR color gamut video is format-converted to form a 10-bit encoded or 16-bit encoded HDR video.
  • 16-bit coded or 10-bit coded HDR video can show higher contrast and richer colors.
  • the method of directly using the color lookup table to realize HDR video conversion can speed up the completion of HDR video conversion. task, improve the efficiency of HDR video conversion.
  • the embodiment of the present application provides a color gamut mapping device.
  • the device embodiment corresponds to the aforementioned method embodiment.
  • the details in the examples are described one by one, but it should be clear that the device in this embodiment can correspondingly implement all the content in the foregoing method embodiments.
  • the present application provides a color gamut mapping device, the above-mentioned device 200 includes:
  • An acquisition unit 201 configured to acquire a first color gamut diagram, where the first color gamut diagram includes a plurality of color node values of the first color gamut;
  • the processing unit 202 is configured to input the first color gamut map into the trained full convolution model for processing, and output the second color gamut map, the second color gamut map includes color node values corresponding to a plurality of color node values one-to-one , the fully convolutional model includes N convolutional layers with a convolution kernel size of 1 ⁇ 1, N-1 activation functions are interspersed in the N convolutional layers, and N is an integer greater than or equal to 3.
  • the activation function is a ReLU activation function.
  • the training methods of the full convolution model include:
  • the training set includes a plurality of first color gamut map samples and a second color gamut map sample corresponding to each first color gamut map sample;
  • the plurality of first color gamut map samples are video frames in at least one first color gamut video sample
  • the second color gamut image sample is a video frame in the second color gamut video sample corresponding to the first color gamut video sample;
  • the preset loss function is used to describe the L2 loss between the predicted second color gamut map and the second color gamut map sample, and the predicted second color gamut map is obtained by processing the first color gamut map sample with the full convolution model of.
  • the processing unit 202 is further configured to determine a color lookup table between the first color gamut map and the second color gamut map according to the first color gamut map and the second color gamut map.
  • the first color gamut is the SDR color gamut
  • the second color gamut is the HDR color gamut
  • the application method of the color lookup table includes: acquiring the SDR video to be processed
  • FIG. 5 is a schematic diagram of a terminal device provided in an embodiment of the present application.
  • the terminal device 300 provided in this embodiment includes: a memory 302 and a processor 301, the memory 302 is used to store computer programs; the processor 301 is used to The methods described in the above method embodiments are executed when the computer program is called, for example, steps S101 to S103 shown in FIG. 1 .
  • the processor 301 executes the computer program, it realizes the functions of the modules/units in the above-mentioned device embodiments, such as the functions of the acquiring unit 201, the processing unit 202, and the determining unit 203 shown in FIG. 4 .
  • the computer program may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 302 and executed by the processor 301 to complete this Apply.
  • the one or more modules/units may be a series of computer program instruction segments capable of accomplishing specific functions, and the instruction segments are used to describe the execution process of the computer program in the terminal device.
  • FIG. 5 is only an example of a terminal device, and does not constitute a limitation on the terminal device. It may include more or less components than those shown in the figure, or combine certain components, or different components, such as
  • the terminal device may also include an input and output device, a network access device, a bus, and the like.
  • the processor 301 may be a central processing unit (Central Processing Unit, CPU), can also be other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), Application Specific Integrated Circuit (Application Specific Integrated Circuit, ASIC), Field Programmable Gate Array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.
  • a general-purpose processor may be a microprocessor, or the processor may be any conventional processor, or the like.
  • the storage 302 may be an internal storage unit of the terminal device, for example, a hard disk or memory of the terminal device.
  • the memory 302 may also be an external storage device of the terminal device, such as a plug-in hard disk equipped on the terminal device, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, Flash Card (Flash Card), etc. Further, the memory 302 may also include both an internal storage unit of the terminal device and an external storage device.
  • the memory 302 is used to store the computer program and other programs and data required by the terminal device.
  • the memory 302 can also be used to temporarily store data that has been output or will be output.
  • the terminal device provided in this embodiment can execute the foregoing method embodiment, and its implementation principle and technical effect are similar, and details are not repeated here.
  • the embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in the foregoing method embodiment is implemented.
  • the embodiment of the present application further provides a computer program product, which, when the computer program product runs on a terminal device, enables the terminal device to implement the method described in the foregoing method embodiments when executed.
  • An embodiment of the present application further provides a chip system, including a processor, the processor is coupled to a memory, and the processor executes a computer program stored in the memory, so as to implement the method described in the above method embodiment.
  • the chip system may be a single chip, or a chip module composed of multiple chips.
  • the above integrated units are realized in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the procedures in the methods of the above embodiments in the present application can be completed by instructing related hardware through computer programs, and the computer programs can be stored in a computer-readable storage medium.
  • the computer program When executed by a processor, the steps in the above-mentioned various method embodiments can be realized.
  • the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate form.
  • the computer-readable storage medium may at least include: any entity or device capable of carrying computer program codes to a photographing device/terminal device, a recording medium, a computer memory, a read-only memory (Read-Only Memory, ROM), a random access Memory (Random Access Memory, RAM), electrical carrier signals, telecommunication signals, and software distribution media.
  • a photographing device/terminal device a recording medium
  • a computer memory a read-only memory (Read-Only Memory, ROM), a random access Memory (Random Access Memory, RAM), electrical carrier signals, telecommunication signals, and software distribution media.
  • ROM read-only memory
  • RAM random access Memory
  • electrical carrier signals telecommunication signals
  • software distribution media such as U disk, mobile hard disk, magnetic disk or optical disk, etc.
  • computer readable media may not be electrical carrier signals and telecommunication signals under legislation and patent practice.
  • references to "one embodiment” or “some embodiments” or the like in this application means that a particular feature, structure, or characteristic described in connection with the embodiment is included in one or more embodiments of the present application.
  • appearances of the phrases “in one embodiment,” “in some embodiments,” “in other embodiments,” “in other embodiments,” etc. in various places in this specification are not necessarily All refer to the same embodiment, but mean “one or more but not all embodiments” unless specifically stated otherwise.
  • the terms “including”, “comprising”, “having” and variations thereof mean “including but not limited to”, unless specifically stated otherwise.
  • first and second are used for description purposes only, and cannot be interpreted as indicating or implying relative importance or implicitly indicating the quantity of indicated technical features. Thus, the features defined as “first” and “second” may explicitly or implicitly include at least one of these features. It should also be understood that the term “and/or” used in the description of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.
  • connection and “connected” should be understood in a broad sense, for example, it can be mechanical connection or electrical connection; it can be direct connection or through An intermediate medium is indirectly connected, which can be the internal communication of two elements or the interaction relationship between two elements. Unless otherwise clearly defined, those of ordinary skill in the art can understand the above terms in this application according to the specific situation. specific meaning.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

本申请提供一种色域映射方法、装置、终端设备及存储介质，涉及图像处理技术领域，能够提高色域映射过程中图像转换的色准。该色域映射方法包括：获取第一色域图，所述第一色域图包括第一色域的多个颜色节点值；将所述第一色域图输入已训练的全卷积模型中处理，输出第二色域图，所述第二色域图中包括与多个颜色节点值一一对应的颜色节点值，所述全卷积模型包括N个卷积核大小为1×1的卷积层，N个所述卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。

Description

一种色域映射方法、装置、终端设备及存储介质 技术领域
本申请涉及图像处理技术领域，尤其涉及一种色域映射方法、装置、终端设备及存储介质。
 背景技术
在实际应用中，同一个原图像或者原视频在不同的显示设备下呈现出来的效果（例如亮度、色彩、对比度和饱和度等）不同，这是由于不同的显示设备所使用的颜色的区域（即色域）不同，常通过色域映射的方式，将图像的源色域转换到目标色域，以使同一原图像或者原视频在不同显示设备上展现出来的效果尽可能一致。
目前，常使用转换函数来是实现色域映射。即通过测量两个色域之间的标定值，并基于该标定值确定转换函数，然后利用转换函数将源色域中的每一个颜色节点的颜色值转换到目标色域中。由于不同色域颜色空间的颜色分布并不是均匀分布的，因此，对每个颜色节点使用同样的转换函数则会产生一定程度的色偏，导致转换后的图像色准较差。
 技术问题
有鉴于此，本申请提供一种色域映射方法、装置、终端设备及存储介质，能够提高色域映射过程中图像转换的色准。
 技术解决方案
第一方面，本申请提供一种色域映射方法，方法包括：获取第一色域图，所述第一色域图包括第一色域的多个颜色节点值；
将所述第一色域图输入已训练的全卷积模型中处理，输出第二色域图，所述第二色域图中包括与所述多个颜色节点值一一对应的颜色节点值，所述全卷积模型包括N个卷积核大小为1×1的卷积层，N个所述卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
可选地，3≤N≤10。
可选地，所述激活函数为非线性激活函数。
可选地，所述全卷积模型的训练方式包括：利用预设的训练集和预设的损失函数对全卷积初始模型进行迭代训练，得到所述全卷积模型；
所述训练集包括多个第一色域图样本以及每个所述第一色域图样本对应的第二色域图样本；所述多个第一色域图样本是至少一个第一色域视频样本中的视频帧，所述第二色域图样本是与所述第一色域视频样本对应的第二色域视频样本中的视频帧；
所述预设的损失函数用于描述预测的第二色域图和所述第二色域图样本之间的L2损失，所述预测的第二色域图为所述全卷积模型对所述第一色域图样本进行处理得到的。
可选的，所述方法还包括：根据所述第一色域图和所述第二色域图确定所述第一色域和所述第二色域图之间的颜色查找表。
可选地，所述第一色域为SDR色域，所述第二色域为HDR色域，所述颜色查找表的应用方法包括：
获取待处理的SDR视频；
利用所述颜色查找表对所述SDR视频进行颜色映射，得到HDR色域视频；
将所述HDR色域视频的编码格式转换为HDR视频编码格式，得到与所述SDR视频对应的HDR视频。
第二方面，本申请提供一种色域映射装置，包括：
获取单元，用于获取第一色域图，所述第一色域图包括第一色域的多个颜色节点值；
处理单元，用于将所述第一色域图输入已训练的全卷积模型中处理，输出第二色域图，所述第二色域图中包括与所述多个颜色节点值一一对应的颜色节点值，所述全卷积模型包括N个卷积核大小为1×1的卷积层，N个所述卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
可选地，3≤N≤10。
第三方面，本申请提供一种终端设备，包括：存储器和处理器，存储器用于存储计算机程序；处理器用于在调用计算机程序时执行上述第一方面中任一方式所述的方法。
第四方面，本申请提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现如上述第一方面中任一方式所述的方法。
第五方面，本申请实施例提供一种计算机程序产品，当计算机程序产品在处理器上运行时，使得处理器执行上述第一方面中任一方式所述的方法。
 有益效果
本申请所提供的一种色域映射方法、装置、终端设备及存储介质，采用全卷积模型来实现色域映射。该全卷积模型包括卷积核大小为1×1的N个卷积层及穿插设置的N-1个激活函数，使第一色域图中的多个颜色节点值都有各自对应的色域映射函数。从而使得经全卷积模型处理后的第二色域图中包括与第一色域图中的多个颜色节点值一一对应的颜色节点值，避免了由于第一色域图中的多个颜色节点值都使用同样的转换函数而产生色偏的情况，因此，采用本申请提供的色域映射方法可以实现非均匀的色域映射，提高色域映射过程中图像转换的色准。
 附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其他的附图。
图1是本申请一实施例提供的一种色域映射方法的流程示意图；
图2是本申请一实施例提供的一种色域映射方法的全卷积模型的架构图；
图3是本申请一实施例提供的HDR色域和SDR色域表示范围的示意图
图4是本申请一实施例提供的一种色域映射装置的示意图；
图5是本申请一实施例提供的一种终端设备的结构示意图。
 本发明的实施方式
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
目前，色域映射通常使用标定法来实现。即通过测量两个色域之间的标定值，并基于该标定值确定转换函数，然后利用转换函数将源色域中的每一个颜色节点的颜色值转换到目标色域中。由于不同色域颜色空间的颜色分布并不是均匀分布的，因此，对每个颜色节点使用同样的转换函数则会产生一定程度的色偏，导致转换后的图像色准较差。
为了提高色域映射过程中图像转换的色准，本申请提供一种色域映射方法，采用全卷积模型来实现色域映射。该全卷积模型包括卷积核大小为1×1的N个卷积层及穿插设置的N-1个激活函数，使第一色域图中的多个颜色节点值都有各自对应的色域映射函数。从而使得经全卷积模型处理后的第二色域图中包括与第一色域图中的多个颜色节点值一一对应的颜色节点值，避免了由于第一色域图中的多个颜色节点值都使用同样的转换函数而产生色偏的情况，因此，采用本申请提供的色域映射方法可以实现非均匀的色域映射，提高色域映射过程中图像转换的色准。
下面以具体地实施例对本申请的技术方案进行详细说明。下面这几个具体的实施例可以相互结合，对于相同或相似的概念或过程可能在某些实施例不再赘述。
如图1所示为本申请提供的一种色域映射方法的一个实施例的流程图。本申请实施例提供的色域映射方法的执行主体可以是图像/视频处理设备，其中，图像/视频处理设备可以是智能手机、平板电脑、摄像机等移动终端设备，也可以是台式电脑、机器人、服务器等能够处理图像/视频数据的终端设备。
如图1所示，本申请实施例中色域映射方法包括：
S101，获取第一色域图，第一色域图包括第一色域的多个颜色节点值。
示例性的，第一色域可以是BT.709色域、BT.2020色域或者DCI-P3色域等，第一色域图可是是由第一色域中的多个颜色节点值构成的图像。
例如，第一色域图可以是拍摄、下载或者是从本地存储区域中读取的颜色节点值在第一色域的图像，也可以是从颜色节点值在第一色域的视频中截取的视频帧或者图像。当然，第一色域图还可以是合成的图像，例如，获取BT.709色域中的全部颜色节点值，利用图像合成工具将上述获取到的BT.709色域中的全部颜色节点值合成图像。
S102，将第一色域图输入已训练的全卷积模型中处理，输出第二色域图，第二色域图中包括与第一色域的多个颜色节点值一一对应的颜色节点值，全卷积模型包括N个卷积核大小为1×1的卷积层，N个卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
其中，N的设置可以根据实际精度要求来设置。例如，N为大于或者等于3的整数。
可选地，激活函数可以是线性整流函数（Rectified Linear Unit，ReLU）。选择ReLU激活函数作为全卷积模型中的激活函数不仅可以加快计算效率，还可以增加全卷积模型的非线性拟合能力。当然，激活函数还可以是其他类型的激活函数，例如，Sigmoid函数或者Thnh函数等。
如图2所示本实施例采用全卷积模型包括3个卷积核大小为1×1的卷积层，及3个卷积层中穿插设置有2个ReLU激活函数。将第一色域图像输入如图2所示的全卷积模型进行处理，能够输出得到第二色域图。
可见，本申请提供的全卷积模型，由于是由卷积核大小为1×1的N个卷积层及穿插设置的N-1个激活函数构成，使第一色域图中的多个颜色节点值都有各自对应的色域映射函数，从而使得经全卷积模型处理后的第二色域图中包括与第一色域图中的多个颜色节点值一一对应的颜色节点值，避免了由于第一色域图中的多个颜色节点值都使用同样的转换函数而产生色偏的情况，因此，采用本申请提供的色域映射方法可以提高色域映射过程中图像转换的色准。
进一步地，由于全卷积模型由卷积核大小为1×1的N个卷积层及穿插设置的N-1个激活函数构成，模型结构简单且使用的参数量相对较少，能够有效降低全卷积模型处理任务的计算成本，提高了计算效率，加快了任务处理的速度。
可以理解的是，由于本申请提供的全卷积模型计算效率高，计算成本小，因此可以直接设置在设备中用来实时处理色域转换任务。
此外，还可以利用本申请提供的全卷积模型制作颜色查找表（color lookup table）。即根据第一色域图和利用全卷积模型生成的第二色域图，确定第一色域和第二色域图之间的颜色查找表。
值得说明的是，颜色查找表可以直接加入相机等终端设备的后期处理过程，以从软件的角度提升相机等终端设备拍摄的图像或视频的质量。颜色查找表也可以应用于图像/视频编辑器中，作为一种图像或者视频后期处理的手段，提高色域映射过程中图像或视频转换的色准，例如，利用颜色查找表实现图像的不同滤镜效果。当然，颜色查找表还可以用于不同显示设备之间的色域映射，使同一图像/视频在不同显示设备上展现出尽可能相同的效果。
下面对全卷积模型的训练方式进行说明。
在本申请中利用预设的训练集和预设的损失函数对全卷积初始模型进行迭代训练，得到全卷积模型。
其中，训练集包括多个第一色域图样本以及每个第一色域图样本对应的第二色域图样本。
在本申请实施例中，第一色域图样本的获取方式（或者来源）可以是通过视频或者图像采集设备直接获取的颜色节点值在第一色域的图像数据，也可以是将获取的颜色节点值在第一色域的视频数据进行抽帧或者拆帧等方式转换成的图像数据。
第二色域图样本也可以是通过视频或者图像采集设备直接获取的颜色节点值在第二色域的图像数据，还可以是将获取的颜色节点值在第二色域的视频数据进行抽帧或者拆帧等方式转换成的图像数据。
在一个示例中，多个第一色域图样本是至少一个第一色域视频样本中的视频帧，第二色域图样本是与第一色域视频样本对应的第二色域视频样本中的视频帧。
不难理解的，相较于单张的图像，视频中包含的颜色节点值更多样，因此，训练集采用第一色域视频样本和对应的第二色域视频样本中够提取到更丰富颜色节点值，采用颜色节点值更丰富的视频帧（包括第一色域图像样本和第二色域图像样本）对全卷积模型进行训练，还能够提高模型训练的准确度。
预设的损失函数用于描述预测的第二色域图和第二色域图样本之间的L2损失，预测的第二色域图为全卷积模型对第一色域图样本进行处理得到的。
值得说明的是，当色域映射的任务不同时，可以通过设计对应的训练集和损失函数来训练初始模型，从而得到适用不同色域映射的任务的全卷积模型。
下面以第一色域为SDR色域，第二色域为HDR色域为例，对本申请提供的全卷积模型的训练过程及应用进行示例性的说明。
如图3所示为HDR色域和SDR色域表示范围的示意图，其中，BT.709和BT.2020都是ITU（国际电信联盟）发布的电视参数标准，DCI-P3是美国电影工业为数字电影院所制定的色域标准。从图3可以看出，DCI-P3、BT.709和BT.2020中范围最大的是BT.2020，DCI-P3的色域范围次之，BT.709所表示的色域范围最小。目前，SDR图像/视频采用的是BT.709色域，而HDR图像/视频采用的是色域范围更为宽广的BT.2020色域或DCI-P3色域。就同一图像/视频而言，无论HDR图像/视频采用BT.2020色域还是DCI-P3色域，HDR图像/视频可以比SDR图像/视频展现出更高的对比度以及更加丰富的色彩。
首先，获取训练集，训练集可以包括多个SDR视频帧样本及与多个SDR视频帧样本一一对应的HDR视频帧样本。
具体的，首先获取SDR视频样本及其对应的HDR视频样本。示例性的，可以从公开的视频网站中获取SDR视频样本及对应的HDR视频样本。也可以对同一RAW数据格式的视频分别进行SDR和HDR处理，得到SDR视频样本及其对应的HDR视频样本。还可以分别利用SDR相机和HDR相机在同一场景下，分别拍摄对应的SDR视频样本和HDR视频样本。
在获取到SDR视频样本及其对应的HDR视频样本之后，分别对SDR视频样本及其对应的HDR视频样本进行抽帧处理，得到多个SDR视频帧样本以及在时序上和空间上与多个SDR视频帧样本一一对应的HDR视频帧样本。
不难理解的，可以采用抽帧工具对SDR视频样本及其对应的HDR视频样本进行抽帧。例如，采用FFmpeg（Fast Forward Mpeg）工具对SDR视频样本及其对应的HDR视频样本进行抽帧。
然后，利用预设的训练集和预设的损失函数对全卷积初始模型进行迭代训练，得到全卷积模型。
在全卷积初始模型搭建完成后，将多个SDR视频帧样本输入到全卷积初始模型中，全卷积初始模型分别对多个SDR视频帧样本进行处理，得到预测的多个HDR视频帧。
预设的损失函数用于全卷积初始模型预测的多个HDR视频帧与HDR视频帧样本之间的L2损失。
基于训练集和上述预设的损失函数，可以通过梯度下降法对全卷积初始模型进行迭代训练，直到模型收敛，即可得到已训练的全卷积模型。
得到具备SDR色域到HDR色域映射的全卷积模型后，可以基于该全卷积模型得到颜色映射表。例如，根据SDR色域中的所有颜色节点值生成SDR色域图，输入至该全卷积模型中处理后，即可得到对应的HDR色域图。然后将SDR色域图和HDR色域图中的颜色节点建立对应关系，得到颜色查找表。
在一个示例中，基于SDR色域图和HDR色域图之间的颜色查找表，本申请实施例提供一种HDR视频转换方法，该方法包括：
首先，获取待处理的SDR视频。
其中，待处理的SDR视频的获取方式可以是拍摄、下载或者是从本地存储区域中读取的完整的视频，也可以是从完成的视频中截取的SDR视频片段，还可以是利用不同色域内的全部或者部分颜色节点值合成不同的色域图。
然后，利用颜色查找表对SDR视频进行颜色映射，得到HDR色域视频。
需要说明的是，待处理的SDR视频采用的是8比特编码格式，经颜色查找表进行颜色映射后的到的HDR色域视频也采用的是8比特编码格式。
最后，将HDR色域的编码格式转换为HDR视频的编码格式，得到与SDR视频对应的HDR视频。
需要说明的是，将8比特编码的HDR色域视频进行格式转换，转换后形成10比特编码或者16比特编码的HDR视频。相较于8比特编码的颜色更新后的SDR视频，16比特编码或者10比特编码的HDR视频能够展现出更高的对比度以及更加丰富的色彩。
值得说明的是，相比于使用SDR视频帧转换为HDR视频帧，再将HDR视频帧合帧处理为HDR视频的转换方法，直接利用颜色查找表实现HDR视频转换的方法能够加快完成HDR视频转换任务，提高HDR视频转换的效率。
基于同一发明构思，作为对上述方法的实现，本申请实施例提供了一种色域映射装置，该装置实施例与前述方法实施例对应，为便于阅读，本装置实施例不再对前述方法实施例中的细节内容进行逐一赘述，但应当明确，本实施例中的装置能够对应实现前述方法实施例中的全部内容。
如图4所示，本申请提供一种色域映射装置，上述装置200包括：
获取单元201，用于获取第一色域图，第一色域图包括第一色域的多个颜色节点值；
处理单元202，用于将第一色域图输入已训练的全卷积模型中处理，输出第二色域图，第二色域图中包括与多个颜色节点值一一对应的颜色节点值，全卷积模型包括N个卷积核大小为1×1的卷积层，N个卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
可选地，3≤N≤10。
可选地，激活函数为ReLU激活函数。
可选地，全卷积模型的训练方式包括：
利用预设的训练集和预设的损失函数对全卷积初始模型进行迭代训练，得到全卷积模型；
训练集包括多个第一色域图样本以及每个第一色域图样本对应的第二色域图样本；多个第一色域图样本是至少一个第一色域视频样本中的视频帧，第二色域图样本是与第一色域视频样本对应的第二色域视频样本中的视频帧；
预设的损失函数用于描述预测的第二色域图和第二色域图样本之间的L2损失，预测的第二色域图为全卷积模型对第一色域图样本进行处理得到的。
可选地，所述处理单元202还用于根据第一色域图和第二色域图确定第一色域和第二色域图之间的颜色查找表。
可选地，第一色域为SDR色域，第二色域为HDR色域，颜色查找表的应用方法包括：获取待处理的SDR视频；
利用颜色查找表对SDR视频进行颜色映射，得到HDR色域视频；将HDR色域视频的编码格式转换为HDR视频的编码格式，得到与SDR视频对应的HDR视频。
所属领域的技术人员可以清楚地了解到，为了描述的方便和简洁，仅以上述各功能单元、模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能单元、模块完成，即将所述装置的内部结构划分成不同的功能单元或模块，以完成以上描述的全部或者部分功能。实施例中的各功能单元、模块可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中，上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。另外，各功能单元、模块的具体名称也只是为了便于相互区分，并不用于限制本申请的保护范围。上述系统中单元、模块的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
基于同一发明构思，本申请实施例还提供了一种终端设备。图5为本申请实施例提供的终端设备的示意图，如图5所示，本实施例提供的终端设备300包括：存储器302和处理器301，存储器302用于存储计算机程序；处理器301用于在调用计算机程序时执行上述方法实施例所述的方法，例如图1所示的步骤S101至步骤S103。或者，所述处理器301执行所述计算机程序时实现上述各装置实施例中各模块/单元的功能，例如图4所示获取单元201、处理单元202及确定单元203的功能。
示例性的，所述计算机程序可以被分割成一个或多个模块/单元，所述一个或者多个模块/单元被存储在所述存储器302中，并由所述处理器301执行，以完成本申请。所述一个或多个模块/单元可以是能够完成特定功能的一系列计算机程序指令段，该指令段用于描述所述计算机程序在所述终端设备中的执行过程。
本领域技术人员可以理解，图5仅仅是终端设备的示例，并不构成对终端设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件，例如所述终端设备还可以包括输入输出设备、网络接入设备、总线等。
所述处理器301可以是中央处理单元（Central Processing Unit，CPU），还可以是其它通用处理器、数字信号处理器（Digital Signal Processor，DSP）、专用集成电路（Application Specific Integrated Circuit，ASIC）、现场可编程门阵列（Field-Programmable Gate Array，FPGA）或者其它可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
所述存储器302可以是所述终端设备的内部存储单元，例如终端设备的硬盘或内存。所述存储器302也可以是所述终端设备的外部存储设备，例如所述终端设备上配备的插接式硬盘，智能存储卡（Smart Media Card，SMC），安全数字（Secure Digital，SD）卡，闪存卡（Flash Card）等。进一步地，所述存储器302还可以既包括所述终端设备的内部存储单元也包括外部存储设备。所述存储器302用于存储所述计算机程序以及所述终端设备所需的其它程序和数据。所述存储器302还可以用于暂时地存储已经输出或者将要输出的数据。
本实施例提供的终端设备可以执行上述方法实施例，其实现原理与技术效果类似，此处不再赘述。
本申请实施例还提供一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现上述方法实施例所述的方法。
本申请实施例还提供一种计算机程序产品，当计算机程序产品在终端设备上运行时，使得终端设备执行时实现上述方法实施例所述的方法。
本申请实施例还提供一种芯片系统，包括处理器，所述处理器与存储器耦合，所述处理器执行存储器中存储的计算机程序，以实现上述方法实施例所述的方法。其中，所述芯片系统可以为单个芯片，或者多个芯片组成的芯片模组。
上述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请实现上述实施例方法中的全部或部分流程，可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一计算机可读存储介质中，该计算机程序在被处理器执行时，可实现上述各个方法实施例的步骤。其中，所述计算机程序包括计算机程序代码，所述计算机程序代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读存储介质至少可以包括：能够将计算机程序代码携带到拍照装置/终端设备的任何实体或装置、记录介质、计算机存储器、只读存储器（Read-Only Memory ，ROM）、随机存取存储器（Random Access Memory，RAM）、电载波信号、电信信号以及软件分发介质。例如U盘、移动硬盘、磁碟或者光盘等。在某些司法管辖区，根据立法和专利实践，计算机可读介质不可以是电载波信号和电信信号。
在本申请中描述的参考“一个实施例”或“一些实施例”等意味着在本申请的一个或多个实施例中包括结合该实施例描述的特定特征、结构或特点。由此，在本说明书中的不同之处出现的语句“在一个实施例中”、“在一些实施例中”、“在其他一些实施例中”、“在另外一些实施例中”等不是必然都参考相同的实施例，而是意味着“一个或多个但不是所有的实施例”，除非是以其他方式另外特别强调。术语“包括”、“包含”、“具有”及它们的变形都意味着“包括但不限于”，除非是以其他方式另外特别强调。
在本申请的描述中，需要理解的是，术语“第一”、“第二”仅用于描述目的，而不能理解为指示或暗示相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征可以明示或者隐含地包括至少一个该特征。还应当理解，在本申请说明书和所附权利要求书中使用的术语“和/或”是指相关联列出的项中的一个或多个的任何组合以及所有可能组合，并且包括这些组合。
此外，在本申请中，除非另有明确的规定和限定，术语“连接”、“相连”等应做广义理解，例如可以是机械连接，也可以是电连接；可以是直接连接，也可以通过中间媒介间接相连，可以是两个元件内部的连通或两个元件的相互作用关系，除非另有明确的限定、对于本领域的普通技术人员而言，可以根据具体情况理解上述术语在本申请中的具体含义。
以上各实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述各实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分或者全部技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。

Claims (10)

  1. 一种色域映射方法，其特征在于，所述方法包括：
    获取第一色域图，所述第一色域图包括第一色域的多个颜色节点值；
    将所述第一色域图输入已训练的全卷积模型中处理，输出第二色域图，所述第二色域图中包括与所述多个颜色节点值一一对应的颜色节点值，所述全卷积模型包括N个卷积核大小为1×1的卷积层，N个所述卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
  2. 根据权利要求1所述的方法，其特征在于，3≤N≤10。
  3. 根据权利要求1所述的方法，其特征在于，所述激活函数为非线性激活函数。
  4. 根据权利要求1所述的方法，其特征在于，所述全卷积模型的训练方式包括：
    利用预设的训练集和预设的损失函数对全卷积初始模型进行迭代训练，得到所述全卷积模型；
    所述训练集包括多个第一色域图样本以及每个所述第一色域图样本对应的第二色域图样本；所述多个第一色域图样本是至少一个第一色域视频样本中的视频帧，所述第二色域图样本是与所述第一色域视频样本对应的第二色域视频样本中的视频帧；
    所述预设的损失函数用于描述预测的第二色域图和所述第二色域图样本之间的L2损失，所述预测的第二色域图为所述全卷积模型对所述第一色域图样本进行处理得到的。
  5. 根据权利要求1-4任一项所述的方法，其特征在于，所述方法还包括：
    根据所述第一色域图和所述第二色域图确定所述第一色域和所述第二色域图之间的颜色查找表。
  6. 根据权利要求5所述的方法，其特征在于，所述第一色域为SDR色域，所述第二色域为HDR色域，所述颜色查找表的应用方法包括：
    获取待处理的SDR视频；
    利用所述颜色查找表对所述SDR视频进行颜色映射，得到HDR色域视频；
    将所述HDR色域视频的编码格式转换为HDR视频编码格式，得到与所述SDR视频对应的HDR视频。
  7. 一种色域映射装置，其特征在于，包括：
    获取单元，用于获取第一色域图，所述第一色域图包括第一色域的多个颜色节点值；
    处理单元，用于将所述第一色域图输入已训练的全卷积模型中处理，输出第二色域图，所述第二色域图中包括与所述多个颜色节点值一一对应的颜色节点值，所述全卷积模型包括N个卷积核大小为1×1的卷积层，N个所述卷积层中穿插设置有N-1个激活函数，N为大于或者等于3的整数。
  8. 根据权利要求7所述的装置，其特征在于，3≤N≤10。
  9. 一种终端设备，包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述计算机程序时实现如权利要求1至6任一项所述的方法。
  10. 一种计算机可读存储介质，所述计算机可读存储介质存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至6任一项所述的方法。
PCT/CN2021/138136 2021-08-02 2021-12-15 一种色域映射方法、装置、终端设备及存储介质 WO2023010753A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110883146.5A CN113781322A (zh) 2021-08-02 2021-08-02 一种色域映射方法、装置、终端设备及存储介质
CN202110883146.5 2021-08-02

Publications (1)

Publication Number Publication Date
WO2023010753A1 true WO2023010753A1 (zh) 2023-02-09

Family

ID=78836589

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/138136 WO2023010753A1 (zh) 2021-08-02 2021-12-15 一种色域映射方法、装置、终端设备及存储介质

Country Status (2)

Country Link
CN (1) CN113781322A (zh)
WO (1) WO2023010753A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781322A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种色域映射方法、装置、终端设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110138259A1 (en) * 2009-12-03 2011-06-09 Microsoft Corporation High Performance Digital Signal Processing In Software Radios
CN113784175A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种hdr视频转换方法、装置、设备及计算机存储介质
CN113781318A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种图像颜色映射方法、装置、终端设备及存储介质
CN113781322A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种色域映射方法、装置、终端设备及存储介质

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102388612B (zh) * 2009-03-13 2014-10-08 杜比实验室特许公司 高动态范围、可视动态范围和宽色域视频的分层压缩
EP3484137A1 (en) * 2017-11-09 2019-05-15 Thomson Licensing Method and device for generating a second image from a first image
CN109903219B (zh) * 2019-02-28 2023-06-30 深圳市商汤科技有限公司 图像处理方法及装置、电子设备、计算机可读存储介质
CN109903256B (zh) * 2019-03-07 2021-08-20 京东方科技集团股份有限公司 模型训练方法、色差校正方法、装置、介质和电子设备
CN110717868B (zh) * 2019-09-06 2022-05-03 上海交通大学 视频高动态范围反色调映射模型构建、映射方法及装置

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110138259A1 (en) * 2009-12-03 2011-06-09 Microsoft Corporation High Performance Digital Signal Processing In Software Radios
CN113784175A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种hdr视频转换方法、装置、设备及计算机存储介质
CN113781318A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种图像颜色映射方法、装置、终端设备及存储介质
CN113781322A (zh) * 2021-08-02 2021-12-10 中国科学院深圳先进技术研究院 一种色域映射方法、装置、终端设备及存储介质

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
HIRAO KATSUHIKO; CHENG ZHENGXUE; TAKEUCHI MASARU; KATTO JIRO: "Convolutional Neural Network based Inverse Tone Mapping for High Dynamic Range Display using LUCORE", 2019 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS (ICCE), 11 January 2019 (2019-01-11), pages 1 - 2, XP033528172, DOI: 10.1109/ICCE.2019.8662048 *
HIRAO KATSUHIKO; CHENG ZHENGXUE; TAKEUCHI MASARU; KATTO JIRO: "Deep Inverse Tone Mapping Optimized for High Dynamic Range Display", 2018 INTERNATIONAL CONFERENCE ON INFORMATION AND COMMUNICATION TECHNOLOGY CONVERGENCE (ICTC), 17 October 2018 (2018-10-17), pages 472 - 474, XP033448093, DOI: 10.1109/ICTC.2018.8539648 *

Also Published As

Publication number Publication date
CN113781322A (zh) 2021-12-10

Similar Documents

Publication Publication Date Title
RU2762384C1 (ru) Переформирование сигналов для сигналов широкого динамического диапазона
WO2023010749A1 (zh) 一种hdr视频转换方法、装置、设备及计算机存储介质
WO2023010754A1 (zh) 一种图像处理方法、装置、终端设备及存储介质
JP7359521B2 (ja) 画像処理方法および装置
WO2023010750A1 (zh) 一种图像颜色映射方法、装置、终端设备及存储介质
US20110205389A1 (en) Methods and Systems for Automatic White Balance
AU2016212243A1 (en) A method and apparatus of encoding and decoding a color picture
CN113810641B (zh) 视频处理方法、装置、电子设备和存储介质
KR20170115503A (ko) 화상을 역-톤 매핑하기 위한 장치 및 방법
WO2020215180A1 (zh) 图像处理方法、装置和电子设备
CN113810642B (zh) 视频处理方法、装置、电子设备和存储介质
JP2004064434A (ja) 画像処理システム、画像撮影装置、画像処理端末、画像処理データ蓄積端末、装置用プログラム、端末用プログラム及び画像データのデータ構造、並びに画像処理方法及び画像データ生成方法
WO2024027287A9 (zh) 图像处理系统及方法、计算机可读介质和电子设备
CN116485979B (zh) 映射关系计算方法、颜色校准方法及电子设备
US20180197282A1 (en) Method and device for producing a digital image
WO2023010755A1 (zh) 一种hdr视频转换方法、装置、设备及计算机存储介质
WO2023010751A1 (zh) 图像高亮区域的信息补偿方法、装置、设备及存储介质
CN110730339A (zh) 一种sdr视频信号处理方法、装置和视频编码设备
CN110807735A (zh) 图像处理方法、装置、终端设备及计算机可读存储介质
WO2023010753A1 (zh) 一种色域映射方法、装置、终端设备及存储介质
US20080068472A1 (en) Digital camera and method
CN113066020A (zh) 图像处理方法及装置、计算机可读介质和电子设备
US10757426B2 (en) Method and apparatus for processing image data
WO2023016044A1 (zh) 视频处理方法、装置、电子设备和存储介质
WO2023016040A1 (zh) 视频处理方法、装置、电子设备和存储介质

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: 21952616

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE