WO2021088333A1 - 图像存储方法、图像读取方法、图像存储器及存储介质 - Google Patents

图像存储方法、图像读取方法、图像存储器及存储介质 Download PDF

Info

Publication number
WO2021088333A1
WO2021088333A1 PCT/CN2020/087161 CN2020087161W WO2021088333A1 WO 2021088333 A1 WO2021088333 A1 WO 2021088333A1 CN 2020087161 W CN2020087161 W CN 2020087161W WO 2021088333 A1 WO2021088333 A1 WO 2021088333A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
image
metadata
information
sub
Prior art date
Application number
PCT/CN2020/087161
Other languages
English (en)
French (fr)
Inventor
鲍明通
Original Assignee
苏州浪潮智能科技有限公司
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 苏州浪潮智能科技有限公司 filed Critical 苏州浪潮智能科技有限公司
Publication of WO2021088333A1 publication Critical patent/WO2021088333A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/51Indexing; Data structures therefor; Storage structures
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/50Information retrieval; Database structures therefor; File system structures therefor of still image data
    • G06F16/58Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/583Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content

Definitions

  • the present invention relates to the field of image storage technology, and in particular to an image storage method, an image reading method, an image memory, an image storage device, an image reading device, and a computer-readable storage medium.
  • the purpose of the present invention is to provide an image storage method, an image reading method, an image memory, an image storage device, an image reading device, and a computer-readable storage medium, which solves the storage efficiency of the existing image storage method Low, making storage costs remain high.
  • an image storage method including:
  • the combination information of the target image is composed of the attribute information and the target metadata information, and the combination information is stored.
  • the acquiring target metadata information corresponding to the target metadata includes:
  • a metadata table is used to obtain the pre-stored address of the target metadata in the metadata storage area, and the conversion coefficient and the pre-stored address are used to form the target metadata information.
  • the process of establishing the metadata table includes:
  • the metadata table is constructed using each of the pre-stored addresses.
  • the method further includes:
  • the storing the special sub-image in the metadata storage area includes:
  • the special sub-image is compressed and stored in the metadata storage area.
  • the obtaining the attribute information of the target image includes:
  • the attribute information is composed of the location information and the identification information of the target image.
  • the present invention also provides an image reading method, including:
  • the target metadata is constructed and processed by using the attribute information and the target metadata information to obtain the target image.
  • the obtaining the corresponding target metadata from the metadata storage area by using the target metadata information includes:
  • the constructing and processing the target metadata by using the attribute information and the target metadata information to obtain the target image includes:
  • the location information is used to construct and process the target sub-image to obtain the target image.
  • the present invention also provides an image memory, including a processor, a memory and an input and output component, wherein:
  • the input and output components are used to obtain or output a target image
  • the memory includes a metadata storage area, a combined information storage area, and a program storage area; wherein the program storage area is used to store computer programs, the metadata storage area is used to store metadata, and the combined information storage area is Used to store combination information;
  • the processor is configured to execute the computer program to implement the above-mentioned image storage method or the above-mentioned image reading method.
  • the present invention also provides an image storage device, including:
  • the decomposition module is used to obtain a target image, and decompose the target image to obtain a target sub-image;
  • a matching module configured to match the target sub-image with the metadata in the metadata storage area, and determine the target metadata corresponding to the target sub-image
  • An obtaining module configured to obtain target metadata information corresponding to the target metadata and attribute information of the target image
  • the storage module is configured to use the attribute information and the target metadata information to form the combined information of the target image, and store the combined information.
  • the present invention also provides an image reading device, including:
  • An instruction acquisition module configured to acquire a read instruction for reading a target image, and use the read instruction to determine the combination information corresponding to the target image;
  • An analysis module configured to analyze the combined information to obtain attribute information and target metadata information, and use the target metadata information to obtain corresponding target metadata from a plurality of metadata in the metadata storage area;
  • the target image acquisition module is configured to use the attribute information and the target metadata information to construct and process the target metadata to obtain the target image.
  • the present invention also provides a computer-readable storage medium for storing a computer program, wherein the computer program is executed by a processor to implement the above-mentioned image storage method or the above-mentioned image reading method.
  • the invention provides an image storage method, which acquires a target image, decomposes the target image, and obtains a target sub-image.
  • the target sub-image is matched with the metadata in the metadata storage area to determine the target metadata corresponding to the target sub-image.
  • the image storage method decomposes the target image to be stored to obtain the target sub-image, which is matched in the metadata storage area to obtain the target metadata. Combining and storing the target metadata information corresponding to the target metadata and the attribute information of the target image is equivalent to storing the target image.
  • the image storage method does not need to store a large number of similar or identical images, only the combination information corresponding to the target image is stored, thus reducing the waste of the storage unit, increasing the reuse rate of the storage unit, improving the storage efficiency, and reducing the storage at the same time. cost.
  • the present invention also provides an image reading method, corresponding to the above-mentioned image storage method, obtaining a reading instruction for reading a target image, and using the reading instruction to determine the combination information corresponding to the target image.
  • the combined information is analyzed to obtain attribute information and target metadata information, and the target metadata information is used to obtain corresponding target metadata from multiple metadata in the metadata storage area. Use the attribute information and the target metadata information to construct and process the target metadata to obtain the target image.
  • the image reading method first obtains the combined information corresponding to the target image when reading the target image, uses the combined information to obtain the target metadata from the metadata storage area, and uses the attribute information and the target metadata information to convert the target metadata.
  • the data can be combined to get the target image.
  • the image reading method does not need to store a large number of similar or identical images, but only needs to store the combination information corresponding to the target image, and read the combination information when the target image is read. Therefore, the waste of the storage unit is reduced, the multiplexing rate of the storage unit is increased, the storage efficiency is improved, and the storage cost is also reduced.
  • the present invention also provides an image memory, an image storage device, an image reading device, and a computer-readable storage medium, which also have the above-mentioned beneficial effects.
  • FIG. 1 is a flowchart of an image storage method provided by an embodiment of the present invention
  • FIG. 3 is a flow chart of a specific construction processing provided by an embodiment of the present invention.
  • FIG. 4 is a schematic structural diagram of an image storage device provided by an embodiment of the present invention.
  • FIG. 5 is a schematic structural diagram of an image reading device provided by an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of an image memory provided by an embodiment of the present invention.
  • FIG. 1 is a flowchart of an image storage method according to an embodiment of the present invention.
  • the method includes:
  • S101 Acquire a target image, decompose the target image, and obtain a target sub-image.
  • the target image is the image you want to store, and its format can be any picture format such as JPEG, TIFF, RAW, BMP, GIF, or PNG.
  • the shape and proportion of the target image can be arbitrarily chosen, for example, it can be square, or it can be length and width. A rectangle with a ratio of 4:3, or a rectangle with an aspect ratio of 16:9.
  • the content of the target image is not limited in this embodiment. For example, it may be a face image or a flower image.
  • the target image After acquiring the target image, the target image is decomposed to obtain the target sub-image.
  • the specific decomposition method is not limited in this embodiment.
  • a deep learning network may be used to identify and mark the target image, and the target image may be decomposed according to the mark of the deep learning network; or the target image may be decomposed according to a preset format.
  • the deep learning network can be trained in advance. Specifically, multiple deep learning networks can be trained to decompose the target image according to the content of the target image.
  • the content of the target image can be identified first.
  • it is a face image
  • the structure of the deep learning model, the training process, and the content of the training image set are not limited in this embodiment, and can be adjusted or selected according to actual conditions.
  • This embodiment does not limit the number of target sub-images, and the specific number of target sub-images is related to the decomposition method.
  • the number of target sub-images is fixed, for example, fixed to 5; when the deep learning network is used to identify and label the target image and decompose it, it can be determined according to the recognition
  • the result of the labeling determines the number of target sub-images.
  • the result of identifying the label may be affected by the training set image of the deep learning model and the structure of the model itself.
  • the target image can be decomposed into multiple target sub-images under certain circumstances, such as When the target image is a face image, the target image can be decomposed into multiple target sub-images such as the mouth, nose, eyes, top of the head, and cheeks; or when the target image is a five-star red flag, it can be divided into five-star and red Two sub-images.
  • the target image can also be decomposed into a sub-image, that is, the target image is determined as a sub-image, for example, when the target image is a pure red red flag, it can be determined as a target sub-image.
  • S102 Match the target sub-image with the metadata in the metadata storage area, and determine the target metadata corresponding to the target sub-image.
  • the target sub-image is matched with the metadata in the metadata storage area.
  • the metadata is stored in the metadata storage area in advance for matching operations with the target sub-image, and is read accordingly. It is used to construct the target image during the process.
  • the specific form of metadata is not limited in this embodiment. For example, it may be a common image; or it may be information composed of certain information elements, such as information composed of a combination of color, size, and rules. For example, it may be red, one The pixel size and the information formed by rotating 360 degrees around a certain point, that is, the information represents a red circle, and the radius of the circle is not limited; or it can be yellow, the size of one pixel, and it moves forward ten times from a certain point. Second, that is, the information represents a yellow line segment with a length of 10 pixels.
  • the metadata that has passed the matching is determined as the target metadata, that is, the target metadata corresponding to the target sub-image.
  • This embodiment does not limit the specific criteria for determining successful matching. For example, when the metadata is a normal image, the target sub-image can be matched with each metadata in the metadata storage area, and the confidence of each match can be determined.
  • the confidence level corresponding to a certain metadata is greater than the confidence threshold, for example, when greater than 99%, the metadata is determined as the target metadata; or the target sub-image and each metadata in the metadata storage area can be Match, and determine the confidence of each match.
  • the confidence corresponding to multiple metadata is greater than the confidence threshold, for example, when it is greater than 99%, the metadata with the highest confidence among the multiple metadata is determined as the target element data.
  • the metadata is information composed of certain information elements, the specific matching process can refer to the prior art, and this embodiment is not limited, as long as the matching can be achieved.
  • S103 Acquire target metadata information corresponding to the target metadata and attribute information of the target image.
  • the target metadata information is information corresponding to the target metadata, which may include the transformation coefficients corresponding to the target sub-image and the target metadata, and the pre-stored address of the target metadata in the metadata storage area. In addition, it may also include the target element. Data number or id information, etc.
  • the attribute information of the target image includes the identification information of the target image, such as id information or number information, and location information.
  • the location information is used to indicate the position of the target sub-image in the target image. It can also be called a mapping factor, that is, the target sub-image. The mapping position of the image in the target image.
  • the transform coefficient between the target sub-image and the corresponding target metadata can be obtained.
  • the transform coefficient may include other kinds of coefficients, such as scaling factor and rotation. Coefficient, length coefficient, size coefficient or flip coefficient, etc., the target metadata is transformed or manipulated according to the scaling factor, and the target sub-image can be obtained.
  • After obtaining the conversion coefficient use the metadata table to obtain the pre-stored address of the target metadata in the metadata storage area.
  • the pre-stored address is the storage address of the target metadata.
  • the metadata table is used to store the pre-stored address of each metadata for the purpose of storing
  • the target metadata information is generated when the target image is generated, or a pre-stored address is provided during the corresponding target image reading process to obtain the target metadata. Using transformation coefficients and pre-stored addresses, the target metadata information can be composed.
  • the corresponding position information of the target sub-image in the target image can be obtained.
  • the position information can be represented by the position of a specific pixel in the target sub-image, or multiple pixels in the target sub-image. For example, you can select the center point of the target sub-image as a specific pixel, and use the coordinate information of the center point in the target image as the position information of the target sub-image; or you can select multiple pixels on the boundary of the target sub-image , Use the coordinate information of these pixels in the target image as the position information of the target sub-image.
  • the location information and the identification information of the target image are used to form the attribute information.
  • each metadata is stored in the metadata storage area, and the pre-stored address corresponding to each metadata is obtained, and the metadata table is constructed using each pre-stored address.
  • the metadata table may also include the number or id of each metadata.
  • S104 Use the attribute information and the target metadata information to compose the combined information of the target image, and store the combined information.
  • the attribute information and the target metadata information are used to form the combination information corresponding to the target image.
  • the specific combination information generation method is not limited in this embodiment.
  • the subsequent order of the data information forms the combined information of the target image; or when there are multiple target metadata information, the target metadata information is sorted, and the attribute information is placed after the target metadata information to form the combined information. After the combined information is generated, the combined information can be stored in the combined information storage area.
  • the target image to be stored is decomposed to obtain the target sub-image, and the target metadata is obtained by matching in the metadata storage area.
  • Combining and storing the target metadata information corresponding to the target metadata and the attribute information of the target image is equivalent to storing the target image.
  • the image storage method does not need to store a large number of similar or identical images, only the combination information corresponding to the target image is stored, thus reducing the waste of the storage unit, increasing the reuse rate of the storage unit, improving the storage efficiency, and reducing the storage at the same time. cost.
  • the target sub-image obtained after decomposing the target image does not match the metadata in the metadata storage area.
  • the target sub-image is After matching with the metadata in the metadata storage area, it can also include:
  • the special sub-image is the target sub-image that fails to match.
  • the criterion for matching failure corresponds to the criterion for matching success in the foregoing embodiment of the invention. For example, when the criterion for matching success is that the confidence between the target sub-image and the target metadata is greater than 99%, Then the criterion for matching failure can be that the confidence of the target sub-image and any metadata time is less than 99%.
  • the target sub-image that fails to match is determined as a special sub-image, the special sub-image is stored in the metadata storage area, and the pre-stored address corresponding to the special sub-image is obtained at the same time.
  • this embodiment does not limit the storage method of storing special sub-images in the metadata storage area.
  • the special sub-images can be directly stored as metadata in the metadata storage area; or the special sub-images can be analyzed. , Get the information corresponding to the special sub-image, and store the information as metadata in the metadata storage area.
  • the embodiment of the present invention in order to reduce the storage space of the metadata storage area occupied by the special sub-images so as to store more metadata in the metadata storage area, it is preferable in the embodiment of the present invention to compress the special sub-images before storing them. Into the metadata storage area.
  • the metadata table is updated with the pre-stored address.
  • the specific update process is not limited in this embodiment.
  • FIG. 2 is a flowchart of an image reading method provided by an embodiment of the present invention, including:
  • S201 Obtain a reading instruction for reading the target image, and use the reading instruction to determine the combination information corresponding to the target image.
  • the read instruction is used to specify the target image to be read, which may include identification information of the target image, or may include number information of combined information, and the like. After obtaining the read instruction, the read instruction is used to determine the combination information corresponding to the target image.
  • the combination information may include attribute information and target metadata information.
  • the attribute information may include identification information of the target image, and the identification information and the reading instruction may be used to determine the combination information corresponding to the target image.
  • S202 Analyze the combined information to obtain attribute information and target metadata information, and use the target metadata information to obtain corresponding target metadata from multiple metadata in the metadata storage area.
  • the attribute information of the target image includes the identification information of the target image, such as id information or number information, and also includes location information.
  • the location information is used to indicate the location of the target sub-image in the target image.
  • the target metadata information is information corresponding to the target metadata, which may include the transformation coefficients corresponding to the target sub-image and the target metadata, and the pre-stored address of the target metadata in the metadata storage area. In addition, it may also include the target element. Data number or id information, etc.
  • the target metadata information is used to obtain corresponding target metadata from a plurality of metadata in the metadata storage area.
  • This embodiment does not limit the number of target metadata information in one combination information.
  • one combination information may include only one target metadata information, or one combination information may include multiple target metadata information.
  • the target metadata information corresponds to the target metadata, so the target metadata information can be used to determine and obtain the target metadata corresponding to the target metadata information from a plurality of metadata in the metadata storage area.
  • the target metadata information when using the target metadata information to obtain the corresponding target metadata from the metadata storage area, the target metadata information can be parsed to obtain the pre-stored address corresponding to the target metadata information, that is, the pre-stored address of the target metadata, Use the pre-stored address to obtain the corresponding target metadata from the metadata storage area.
  • S203 Use the attribute information and the target metadata information to construct and process the target metadata to obtain the target image.
  • the construction process can specifically include transformation processing and combination processing.
  • the specific process is not limited in this embodiment.
  • the target metadata information is used to transform the target metadata.
  • Applying the image reading method provided by the embodiment of the present invention when reading the target image, first obtain the combination information corresponding to the target image, use the combination information to obtain the target metadata from the metadata storage area, and use the attribute information and the target element
  • the data information combines the target metadata to obtain the target image.
  • the image reading method does not need to store a large number of similar or identical images, but only needs to store the combination information corresponding to the target image, and read the combination information when reading the target image. Therefore, the waste of the storage unit is reduced, the multiplexing rate of the storage unit is increased, the storage efficiency is improved, and the storage cost is also reduced.
  • FIG. 3 is a specific construction process flowchart provided by an embodiment of the present invention, including:
  • the target metadata information includes the transformation coefficient and the pre-stored address of the target metadata in the metadata storage area.
  • the transformation coefficient is used to transform the target metadata into the corresponding target sub-image, and the target sub-image is obtained by decomposing the target image.
  • the corresponding target metadata is processed using the transform coefficient to obtain the corresponding target sub-image.
  • the transform coefficient may include the f parameter and the m parameter.
  • the f parameter and the m parameter are used to perform ⁇ a-s-m ⁇ *m*f calculation on the target metadata to obtain the target sub-image.
  • the * operator only represents an operation, and the specific operation method and content are not limited in this embodiment.
  • the attribute information includes position information of the target sub-image in the target image and identification information of the target image.
  • the position information is used to indicate the position of the target sub-image in the target image.
  • S304 Use the position information to combine the target sub-images to obtain the target image.
  • the target sub-images can be combined by the position to obtain the target image. It should be noted that the embodiment of the present invention only describes a specific construction process, and other similar methods can also be used in the image reading method provided by the present invention.
  • the image storage device provided by the embodiment of the present invention will be introduced below.
  • the image storage device described below and the image storage method described above may correspond to each other and refer to each other.
  • FIG. 4 is a schematic structural diagram of an image storage device according to an embodiment of the present invention, including:
  • the decomposition module 410 is used to obtain a target image, and decompose the target image to obtain a target sub-image;
  • the matching module 420 is configured to match the target sub-image with the metadata in the metadata storage area, and determine the target metadata corresponding to the target sub-image;
  • the obtaining module 430 is configured to obtain target metadata information corresponding to the target metadata and attribute information of the target image;
  • the storage module 440 is configured to use the attribute information and the target metadata information to compose the combined information of the target image, and store the combined information.
  • the obtaining module 430 includes:
  • a transform coefficient acquisition unit which is used to acquire transform coefficients between the target sub-image and the corresponding target metadata
  • the pre-stored address obtaining unit is used to obtain the pre-stored address of the target metadata in the metadata storage area by using the metadata table, and use the conversion coefficient and the pre-stored address to form the target metadata information.
  • Optional include:
  • the template storage module is used to store each metadata in the metadata storage area and obtain the pre-stored address corresponding to each metadata;
  • the metadata table building module is used to construct a metadata table using each pre-stored address.
  • it also includes:
  • the special storage module is used to store the special sub-image in the metadata storage area and obtain the pre-stored address corresponding to the special sub-image; among them, the special sub-image is the target sub-image that fails to match;
  • the metadata table update module is used to update the metadata table with the pre-stored address corresponding to the special sub-image.
  • special storage modules including:
  • the compression unit is used to compress the special sub-images and store them in the metadata storage area.
  • the obtaining module 430 includes:
  • the location information acquiring unit is used to acquire the location information corresponding to the target sub-image in the target image
  • the composition unit is used to compose attribute information by using the location information and the identification information of the target image.
  • the image reading device provided by the embodiment of the present invention will be introduced below.
  • the image reading device described below and the image reading method described above can be referred to each other.
  • FIG. 5 is a schematic structural diagram of an image reading device provided by an embodiment of the present invention, including:
  • the instruction acquisition module 510 is configured to acquire a reading instruction for reading the target image, and use the reading instruction to determine the combination information corresponding to the target image;
  • the parsing module 520 is configured to analyze the combined information to obtain attribute information and target metadata information, and use the target metadata information to obtain corresponding target metadata from multiple metadata in the metadata storage area;
  • the target image acquisition module 530 is used to construct and process target metadata by using attribute information and target metadata information to obtain a target image.
  • the parsing module 520 includes:
  • the pre-stored address resolution unit is used to parse the target metadata information to obtain the corresponding pre-stored address
  • the target metadata obtaining unit is configured to obtain the corresponding target metadata from the metadata storage area by using the pre-stored address.
  • the target image acquisition module 530 includes:
  • the transformation coefficient analysis unit is used to analyze the target metadata information to obtain the transformation coefficient
  • the transformation processing unit is configured to process the corresponding target metadata by using the transformation coefficient to obtain the corresponding target sub-image
  • the location information analysis unit is used to analyze the attribute information to obtain the location information corresponding to the target sub-image
  • the acquiring unit is used to construct and process the target sub-image by using the position information to obtain the target image.
  • the image storage device provided by the embodiment of the present invention will be introduced below.
  • the image storage device described below and the image storage method or image reading method described above may correspond to each other with reference to each other.
  • FIG. 6 is a schematic structural diagram of an image memory provided by an embodiment of the present invention.
  • the image memory includes an input and output component 610, a memory 630, and a processor 620, wherein:
  • the input and output component 610 is used to obtain or output a target image
  • the memory 630 includes a metadata storage area, a combination information storage area, and a program storage area; wherein the program storage area is used to store computer programs, the metadata storage area is used to store metadata, and the combination information storage area is used to store combination information;
  • the processor 620 is configured to execute a computer program to implement the above-mentioned image storage method or the above-mentioned image reading method.
  • the computer-readable storage medium provided by the embodiments of the present invention will be introduced below.
  • the computer-readable storage medium described below and the image storage method or image reading method described above can be referred to each other correspondingly.
  • the present invention also provides a computer-readable storage medium on which a computer program is stored.
  • the computer program is executed by a processor, the steps of the above-mentioned image storage method or the above-mentioned image reading method are realized.
  • the computer-readable storage medium may include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk, etc., which can store program codes Medium.
  • the steps of the method or algorithm described in combination with the embodiments disclosed in this document can be directly implemented by hardware, a software module executed by a processor, or a combination of the two.
  • the software module can be placed in random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or all areas in the technical field. Any other known storage media.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Library & Information Science (AREA)
  • Software Systems (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种图像存储方法,包括:获取目标图像,对目标图像进行分解,得到目标子图像;将目标子图像与元数据存储区中的元数据进行匹配,确定目标子图像对应的目标元数据;获取目标元数据对应的目标元数据信息和目标图像的属性信息;利用属性信息和目标元数据信息组成目标图像的组合信息,并存储组合信息;该方法将目标元数据对应的目标元数据信息和目标图像的属性信息进行组合并存储,相当于将目标图像存储了起来,无需存储大量相同的图像,仅需存储目标图像对应的组合信息,减少了存储单元的浪费;本发明还提供了一种图像读取方法、图像存储装置、图像读取装置、图像存储器及计算机可读存储介质,同样具有上述有益效果。

Description

图像存储方法、图像读取方法、图像存储器及存储介质
本申请要求于2019年11月07日提交中国专利局、申请号为201911082402.X、发明名称为“图像存储方法、图像读取方法、图像存储器及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本发明涉及图像存储技术领域,特别涉及一种图像存储方法、图像读取方法、图像存储器、图像存储装置、图像读取装置及计算机可读存储介质。
背景技术
在电子信息技术快速发展的今天,特别是随着人工智能、大数据分析、AR技术的发展,需要存储的图像越来越多。由于某一个存储集群中存储的图像大多为同一领域,例如为人脸图像,或者为花卉图像,因此存储集群中很多图像十分相似,导致存储了很多相近或相同的图像,这造成了大量存储单元的浪费,存储单元复用率不高,进而使得存储效率较低,同时还使得存储成本居高不下。
因此,如何解决现有图像存储方法存在的存储效率较低,使得存储成本居高不下的问题,是本领域技术人员需要解决的技术问题。
发明内容
有鉴于此,本发明的目的在于提供一种图像存储方法、图像读取方法、图像存储器、图像存储装置、图像读取装置及计算机可读存储介质,解决了现有图像存储方法存在的存储效率较低,使得存储成本居高不下的问题。
为解决上述技术问题,本发明提供了一种图像存储方法,包括:
获取目标图像,对所述目标图像进行分解,得到目标子图像;
将所述目标子图像与元数据存储区中的元数据进行匹配,确定所述目标子图像对应的目标元数据;
获取所述目标元数据对应的目标元数据信息和所述目标图像的属性信息;
利用所述属性信息和所述目标元数据信息组成所述目标图像的组合信息,并存储所述组合信息。
可选的,所述获取所述目标元数据对应的目标元数据信息,包括:
获取所述目标子图像与对应的所述目标元数据之间的变换系数;
利用元数据表获取所述目标元数据在所述元数据存储区中的预存地址,利用所述变换系数和所述预存地址组成所述目标元数据信息。
可选的,所述元数据表的建立过程,包括:
将各个所述元数据存入所述元数据存储区,并获取各个所述元数据对应的所述预存地址;
利用各个所述预存地址构建所述元数据表。
可选的,在所述将所述目标子图像与元数据存储区中的元数据进行匹配之后,还包括:
将特殊子图像存入所述元数据存储区,并获取所述特殊子图像对应的预存地址;其中,所述特殊子图像为匹配失败的目标子图像;
利用所述特殊子图像对应的所述预存地址更新所述元数据表。
可选的,所述将特殊子图像存入所述元数据存储区,包括:
对所述特殊子图像进行压缩处理后存入所述元数据存储区。
可选的,所述获取所述目标图像的属性信息,包括:
获取所述目标子图像在所述目标图像中对应的位置信息;
利用所述位置信息和所述目标图像的标识信息组成所述属性信息。
本发明还提供了一种图像读取方法,包括:
获取用于读取目标图像的读取指令,利用所述读取指令确定所述目标图像对应的组合信息;
对所述组合信息进行解析,得到属性信息和目标元数据信息,利用所述目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据;
利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建 处理,得到所述目标图像。
可选的,所述利用所述目标元数据信息从元数据存储区中获取对应的目标元数据,包括:
对所述目标元数据信息进行解析,得到对应的预存地址;
利用所述预存地址从所述元数据存储区中获取对应的所述目标元数据。
可选的,所述利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建处理,得到所述目标图像,包括:
对所述目标元数据信息进行解析,得到变换系数;
利用所述变换系数对对应的所述目标元数据进行处理,得到对应的目标子图像;
对所述属性信息进行解析,得到所述目标子图像对应的位置信息;
利用所述位置信息对所述目标子图像进行构建处理,得到所述目标图像。
本发明还提供了一种图像存储器,包括处理器、存储器和输入输出部件,其中:
所述输入输出部件,用于获取或输出目标图像;
所述存储器,包括元数据存储区、组合信息存储区和程序存储区;其中,所述程序存储区用于保存计算机程序,所述元数据存储区用于保存元数据,所述组合信息存储区用于存储组合信息;
所述处理器,用于执行所述计算机程序,以实现上述的图像存储方法或上述的图像读取方法。
本发明还提供了一种图像存储装置,包括:
分解模块,用于获取目标图像,对所述目标图像进行分解,得到目标子图像;
匹配模块,用于将所述目标子图像与元数据存储区中的元数据进行匹配,确定所述目标子图像对应的目标元数据;
获取模块,用于获取所述目标元数据对应的目标元数据信息和所述目标图像的属性信息;
存储模块,用于利用所述属性信息和所述目标元数据信息组成所述目标图像的组合信息,并存储所述组合信息。
本发明还提供了一种图像读取装置,包括:
指令获取模块,用于获取用于读取目标图像的读取指令,利用所述读取指令确定所述目标图像对应的组合信息;
解析模块,用于对所述组合信息进行解析,得到属性信息和目标元数据信息,利用所述目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据;
目标图像获取模块,用于利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建处理,得到所述目标图像。
本发明还提供了一种计算机可读存储介质,用于保存计算机程序,其中,所述计算机程序被处理器执行时实现上述的图像存储方法或上述的图像读取方法。
本发明提供了一种图像存储方法,获取目标图像,对目标图像进行分解,得到目标子图像。将目标子图像与元数据存储区中的元数据进行匹配,确定目标子图像对应的目标元数据。获取目标元数据对应的目标元数据信息和目标图像的属性信息。利用属性信息和目标元数据信息组成目标图像的组合信息,并存储组合信息。
可见,该图像存储方法将想要存储的目标图像进行分解,得到目标子图像,并在元数据存储区中匹配得到目标元数据。将目标元数据对应的目标元数据信息和目标图像的属性信息进行组合并存储,即相当于将目标图像存储了起来。该图像存储方法无需存储大量相近或相同的图像,仅需存储目标图像对应的组合信息,因此减少了存储单元的浪费,提高了存储单元的复用率,提高了存储效率,同时还降低了存储成本。
本发明还提供了一种图像读取方法,与上述图像存储方法相对应,获取用于读取目标图像的读取指令,利用读取指令确定目标图像对应的组合信息。对组合信息进行解析,得到属性信息和目标元数据信息,利用目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据。利用属性信息和目标元数据信息对目标元数据进行构建处理,得到目标图像。
可见,该图像读取方法在读取目标图像时先获取目标图像对应的组合信息,利用所述组合信息从元数据存储区中获取目标元数据,并利用属性信息和目标元数据信息将目标元数据进行组合即可得到目标图像。该图像读取方法无需存储大量相近或相同的图像,仅需存储目标图像对应的组合信息,并在读取目标图像时读取组合信息即可。因此减少了存储单元的浪费,提高了存储单元的复用率,提高了存储效率,同时还降低了存储成本。
此外,本发明还提供了一种图像存储器、图像存储装置、图像读取装置及计算机可读存储介质,同样具有上述有益效果。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,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发明的实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据提供的附图获得其他的附图。
图1为本发明实施例提供的一种图像存储方法流程图;
图2为本发明实施例提供的一种图像读取方法流程图;
图3为本发明实施例提供的一种具体的构建处理流程图;
图4为本发明实施例提供的一种图像存储装置的结构示意图;
图5为本发明实施例提供的一种图像读取装置的结构示意图;
图6为本发明实施例提供的一种图像存储器的结构示意图。
具体实施方式
为使本发明实施例的目的、技术方案和优点更加清楚,下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
请参考图1,图1为本发明实施例提供的一种图像存储方法流程图。该 方法包括:
S101:获取目标图像,对目标图像进行分解,得到目标子图像。
目标图像为想要存储的图像,其格式可以为JPEG、TIFF、RAW、BMP、GIF或PNG等任一图片格式,目标图像的形状和比例可以任取,例如可以为正方形,或者可以为长宽比为4:3的长方形,或者可以为长宽比为16:9的长方形。目标图像的内容本实施例不做限定,例如可以为人脸图像、或者可以为花卉图像。
在获取目标图像后,对目标图像进行分解,得到目标子图像。具体的分解方法本实施例不做限定,例如可以利用深度学习网络对目标图像进行识别和标记,按照深度学习网络的标记对目标图像进行分解;或者可以按照预设格式对目标图像进行分解。为了使分解得到的目标子图像与元数据更加相似,本实施例中优选的,利用深度学习网络对目标图像进行识别和标记,并按照标记对目标图像进行分解。深度学习网络可以提前训练好,具体的,可以训练多个深度学习网络,以便根据目标图像的内容对目标图像进行分解,例如可以在获取到目标图像后,先识别目标图像的内容,当目标图像为人脸图像时,利用人脸识别模型对目标图像进行标记,再根据标记对目标图像进行划分;或者当目标图像为花卉图像时,利用花卉识别模型对目标图像进行标记,再根据标记对目标图像进行划分。深度学习模型的结构、训练过程和训练图像集的内容本实施例均不做限定,可以根据实际的情况进行调整或选择。
本实施例并不限定目标子图像的数量多少,目标子图像具体的数量大小与分解方法有关。具体的,当按照预设规则对目标图像进行分解时,则目标子图像的数量是固定的,例如固定为5个;当利用深度学习网络对目标图像进行识别标记并分解时,则可以根据识别标记的结果确定目标子图像的数量,识别标记的结果可能会受到深度学习模型的训练集图像和模型本身结构的影响,例如在某种情况下可以将目标图像分解为多个目标子图像,例如当目标图像为人脸图像时,可以将目标图像分解为嘴部、鼻部、眼部、头顶、脸颊等多个目标子图像;或者当目标图像为五星红旗时,可以将其分为五星和红色两个子图像。当然,目标图像也可以被分解为一个 子图像,即将目标图像确定为一个子图像,例如像目标图像为纯红色的红旗时,则可以将其确定为一个目标子图像。
S102:将目标子图像与元数据存储区中的元数据进行匹配,确定目标子图像对应的目标元数据。
在得到目标子图像后,将目标子图像与元数据存储区中的元数据进行匹配,元数据被提前存入元数据存储区,用于与目标子图像进行匹配操作,并在相应的读取过程中用于构建目标图像。元数据的具体形式本实施例不做限定,例如可以为普通的图像;或者可以为某些信息元素组成的信息,例如为颜色、大小和规则等组合而成的信息,例如可以为红色、一个像素大小、绕某点旋转360度组合而成的信息,即该信息表示了一个红色的圆,圆的半径大小不做限定;或者可以为黄色,一个像素大小,从某点开始向前移动十次,即该信息表示了一个黄色的长度为10像素的线段。
具体的,在将目标子图像与元数据进行匹配之后,将匹配通过的元数据确定为目标元数据,即与目标子图像对应的目标元数据。本实施利并不限定具体的匹配成功的判定标准,例如当元数据为普通图像时,可以将目标子图像与元数据存储区中的每个元数据进行匹配,并确定每次匹配的置信度,当某一元数据对应的置信度大于置信度阈值时,例如当大于99%时,将该元数据确定为目标元数据;或者可以将目标子图像与元数据存储区中的每个元数据进行匹配,并确定每次匹配的置信度,当多个元数据对应的置信度大于置信度阈值时,例如当大于99%时,将这多个元数据中置信度最大的元数据确定为目标元数据。当元数据为某些信息元素组成的信息时,具体的匹配过程可以参考现有技术,本实施例并不限定,只要可以实现匹配即可。
S103:获取目标元数据对应的目标元数据信息和目标图像的属性信息。
目标元数据信息为目标元数据对应的信息,其可以包括目标子图像和目标元数据对应的变换系数以及目标元数据在元数据存储区中的预存地址,除此之外,还可以包括目标元数据的编号或id信息等。当有多个目标子图像时,相应的会有多个目标元数据,每个目标元数据对应一个目标元数据信息,因此会有多个目标元数据信息。目标图像的属性信息包括目标 图像的标识信息,例如id信息或编号信息,还包括位置信息,位置信息用于表示目标子图像在目标图像中的位置,也可以被称为映射因子,即目标子图像在目标图像中的映射位置。
在本发明实施例中,获取目标元数据对应的目标元数据信息时,可以获取目标子图像与对应的目标元数据之间的变换系数,变换系数可以包括其他多种系数,例如缩放系数、旋转系数、长度系数、大小系数或翻转系数等,目标元数据按照缩放系数进行变换或操作,即可得到目标子图像。在获取变换系数后,利用元数据表获取目标元数据在元数据存储区中的预存地址,预存地址为目标元数据的存储地址,元数据表用于存储各个元数据的预存地址,以便在存储目标图像时生成目标元数据信息,或者在相应的目标图像读取过程中提供预存地址以便获取目标元数据。利用变换系数和预存地址,即可组成目标元数据信息。
在获取目标图像的属性信息时,可以获取目标子图像在目标图像中对应的位置信息,该位置信息可以利用目标子图像中一个特定像素的位置进行表示,也可以利用目标子图像中多个像素的位置进行表示,例如可以选择目标子图像的中心点作为特定像素,将该中心点在目标图像中的坐标信息作为目标子图像的位置信息;或者可以选择目标子图像的边界上的多个像素,利用这些像素在目标图像中的坐标信息作为目标子图像的位置信息。在获取位置信息后,利用位置信息和目标图像的标识信息组成属性信息。
进一步,在利用元数据表获取预存地址之前,还需要构建元数据表。具体的,将各个元数据存入元数据存储区中,并获取各个元数据对应的预存地址,利用各个预存地址构建元数据表。元数据表中还可以包括各个元数据的编号或id。
S104:利用属性信息和目标元数据信息组成目标图像的组合信息,并存储组合信息。
在获取属性信息和目标元数据信息之后,利用属性信息和目标元数据信息组成目标图像对应的组合信息,具体的组合信息生成方法本实施例不做限定,例如可以按照属性信息在前,目标元数据信息在后的顺序组成目标图像的组合信息;或者在目标元数据信息有多个时,对目标元数据信息 排序,并将属性信息放置于目标元数据信息的后面组成组合信息。在生成组合信息之后,可以将组合信息存储于组合信息存储区中。
应用本发明实施例提供的图像存储方法,将想要存储的目标图像进行分解,得到目标子图像,并在元数据存储区中匹配得到目标元数据。将目标元数据对应的目标元数据信息和目标图像的属性信息进行组合并存储,即相当于将目标图像存储了起来。该图像存储方法无需存储大量相近或相同的图像,仅需存储目标图像对应的组合信息,因此减少了存储单元的浪费,提高了存储单元的复用率,提高了存储效率,同时还降低了存储成本。
基于上述发明实施例,在一种可能的实施方式中,对目标图像进行分解后得到的目标子图像并不与元数据存储区中的元数据相匹配,为了解决上述问题,在将目标子图像和元数据存储区中的元数据进行匹配之后,还可以包括:
将特殊子图像存入元数据存储区,并获取特殊子图像对应的预存地址。
需要说明的是,特殊子图像即为匹配失败的目标子图像。在本发明实施例中,匹配失败的判定标准与上述发明实施例中匹配成功的判定标准相对应,例如当匹配成功的判定标准为目标子图像与目标元数据之间的置信度大于99%,则匹配失败的判定标准即可为目标子图像与任一元数据时间的置信度均小于99%。在匹配失败后,将匹配失败的目标子图像确定为特殊子图像,并将特殊子图像存入元数据存储区,同时获取特殊子图像对应的预存地址。需要说明的是,本实施例并不限定将特殊子图像存储元数据存储区中的存储方法,例如可以直接将特殊子图像作为元数据存入元数据存储区;或者可以对特殊子图像进行解析,得到与特殊子图像对应的信息,将该信息作为元数据存入元数据存储区。
进一步,为了减少特殊子图像占用元数据存储区的存储空间,以便在元数据存储区中存入更多元数据,本发明实施例中优选的,对特殊子图像进行压缩处理后再将其存入元数据存储区。
利用特殊子图像对应的预存地址更新元数据表。
再将特殊子图像存入元数据存储区并获取对应的预存地址之后,利用 该预存地址更新元数据表。在更新元数据表的过程中,可能还需要对特殊子图像进行编号处理以获取特殊子图像对应的图像编号,并将图像编号加入元数据表,具体的更新过程本实施例不做限定。
本发明实施例将说明一种与上述图像存储方法相对应的图像读取方法,具体请参考图2,图2为本发明实施例提供的一种图像读取方法流程图,包括:
S201:获取用于读取目标图像的读取指令,利用读取指令确定目标图像对应的组合信息。
读取指令用于指定读取的目标图像,其中可以包括目标图像的标识信息,或者可以包括组合信息的编号信息等。在获取读取指令之后,利用读取指令确定目标图像对应的组合信息。组合信息可以包括属性信息和目标元数据信息,属性信息中可以包括目标图像的标识信息,可以利用该标识信息和读取指令确定目标图像对应的组合信息。
S202:对组合信息进行解析,得到属性信息和目标元数据信息,利用目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据。
对组合信息进行解析之后,得到属性信息和目标元数据信息。目标图像的属性信息包括目标图像的标识信息,例如id信息或编号信息,还包括位置信息,位置信息用于表示目标子图像在目标图像中的位置。目标元数据信息为目标元数据对应的信息,其可以包括目标子图像和目标元数据对应的变换系数以及目标元数据在元数据存储区中的预存地址,除此之外,还可以包括目标元数据的编号或id信息等。
在获取目标元数据信息之后,利用目标元数据信息从元数据存储区中的多个元数据中获取对应目标元数据。本实施例并不限定一个组合信息中目标元数据信息的数量,例如一个组合信息可以仅包括一个目标元数据信息,或者一个组合信息中可以包括多个目标元数据信息。目标元数据信息与目标元数据相对应,因此可以利用目标元数据信息从元数据存储区中的多个元数据中确定并获取与目标元数据信息相对应的目标元数据。
具体的,在利用目标元数据信息从元数据存储区中获取对应的目标元数据时,可以对目标元数据信息进行解析,得到目标元数据信息对应的预存地址,即目标元数据的预存地址,利用预存地址从元数据存储区中获取对应的目标元数据。
S203:利用属性信息和目标元数据信息对目标元数据进行构建处理,得到目标图像。
利用属性信息和目标元数据信息对目标图像进行构建处理,即可得到目标图像。构建处理具体可以包括变换处理和组合处理,其具体过程本实施例不做限定,例如当目标元数据信息有多个时,利用目标元数据信息分别对目标元数据进行变换处理,在变换处理后利用属性信息将经过变换处理的目标元数据进行组合,即可得到目标图像。在得到目标图像后,可以执行其他操作,例如可以输出目标图像,或可以展示目标图像。
应用本发明实施例提供的图像读取方法,在读取目标图像时先获取目标图像对应的组合信息,利用所述组合信息从元数据存储区中获取目标元数据,并利用属性信息和目标元数据信息将目标元数据进行组合即可得到目标图像。该图像读取方法无需存储大量相近或相同的图像,仅需存储目标图像对应的组合信息,并在读取目标图像时读取组合信息即可。因此减少了存储单元的浪费,提高了存储单元的复用率,提高了存储效率,同时还降低了存储成本。
基于上述发明实施例,本发明实施例将说明一种具体的构建处理流程,即对S203步骤进行具体说明。请参考图3,图3为本发明实施例提供的一种具体的构建处理流程图,包括:
S301:对目标元数据信息进行解析,得到变换系数。
在本发明实施例中,目标元数据信息包括变换系数和目标元数据在元数据存储区中的预存地址。变换系数用于将目标元数据变换为对应的目标子图像,目标子图像由目标图像分解得到。
S302:利用变换系数对对应的目标元数据进行处理,得到对应的目标子图像。
在得到变换系数后,利用变换系数对对应的目标元数据进行处理,得到对应的目标子图像,本实施例并不限定具体的处理过程。例如,变换系数可以包括f参数和m参数,当目标元数据为{a-s-m}时,利用f参数和m参数对目标元数据进行{a-s-m}*m*f计算,得到目标子图像。在本发明实施例中,*运算符仅表示运算,对于具体的运算方法和内容本实施例不做限定。
S303:对属性信息进行解析,得到目标子图像对应的位置信息。
在本发明实施例中,属性信息包括目标子图像在目标图像中的位置信息和目标图像的标识信息。位置信息用于表示目标子图像在目标图像中的位置。
S304:利用位置信息对目标子图像进行组合,得到目标图像。
在获取位置信息之后,可以利用位置对目标子图像进行组合,以便得到目标图像。需要说明的是,本发明实施例仅说明了一种具体的构建处理过程,其他类似的方法也可以用于本发明提供的图像读取方法中。
下面对本发明实施例提供的图像存储装置进行介绍,下文描述的图像存储装置与上文描述的图像存储方法可相互对应参照。
请参考图4,图4为本发明实施例提供的一种图像存储装置的结构示意图,包括:
分解模块410,用于获取目标图像,对目标图像进行分解,得到目标子图像;
匹配模块420,用于将目标子图像与元数据存储区中的元数据进行匹配,确定目标子图像对应的目标元数据;
获取模块430,用于获取目标元数据对应的目标元数据信息和目标图像的属性信息;
存储模块440,用于利用属性信息和目标元数据信息组成目标图像的组合信息,并存储组合信息。
可选的,获取模块430,包括:
变换系数获取单元,用于获取目标子图像与对应的目标元数据之间的 变换系数;
预存地址获取单元,用于利用元数据表获取目标元数据在元数据存储区中的预存地址,利用变换系数和预存地址组成目标元数据信息。
可选的,包括:
模板存储模块,用于将各个元数据存入元数据存储区,并获取各个元数据对应的预存地址;
元数据表构建模块,用于利用各个预存地址构建元数据表。
可选的,还包括:
特殊存储模块,用于将特殊子图像存入元数据存储区,并获取特殊子图像对应的预存地址;其中,特殊子图像为匹配失败的目标子图像;
元数据表更新模块,用于利用特殊子图像对应的预存地址更新元数据表。
可选的,特殊存储模块,包括:
压缩单元,用于对特殊子图像进行压缩处理后存入元数据存储区。
可选的,获取模块430,包括:
位置信息获取单元,用于获取目标子图像在目标图像中对应的位置信息;
组成单元,用于利用位置信息和目标图像的标识信息组成属性信息。
下面对本发明实施例提供的图像读取装置进行介绍,下文描述的图像读取装置与上文描述的图像读取方法可相互对应参照。
请参考图5,图5为本发明实施例提供的一种图像读取装置的结构示意图,包括:
指令获取模块510,用于获取用于读取目标图像的读取指令,利用读取指令确定目标图像对应的组合信息;
解析模块520,用于对组合信息进行解析,得到属性信息和目标元数据信息,利用目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据;
目标图像获取模块530,用于利用属性信息和目标元数据信息对目标 元数据进行构建处理,得到目标图像。
可选的,解析模块520,包括:
预存地址解析单元,用于对目标元数据信息进行解析,得到对应的预存地址;
目标元数据获取单元,用于利用预存地址从元数据存储区中获取对应的目标元数据。
可选的,目标图像获取模块530,包括:
变换系数解析单元,用于对目标元数据信息进行解析,得到变换系数;
变换处理单元,用于利用变换系数对对应的目标元数据进行处理,得到对应的目标子图像;
位置信息解析单元,用于对属性信息进行解析,得到目标子图像对应的位置信息;
获取单元,用于利用位置信息对目标子图像进行构建处理,得到目标图像。
下面对本发明实施例提供的图像存储设备进行介绍,下文描述的图像存储设备与上文描述的图像存储方法或图像读取方法可相互对应参照。
请参考图6,图6为本发明实施例提供的一种图像存储器的结构示意图,该图像存储器包括输入输出部件610、存储器630和处理器620,其中:
输入输出部件610,用于获取或输出目标图像;
存储器630,包括元数据存储区、组合信息存储区和程序存储区;其中,程序存储区用于保存计算机程序,元数据存储区用于保存元数据,组合信息存储区用于存储组合信息;
处理器620,用于执行计算机程序,以实现上述的图像存储方法或上述的图像读取方法。
下面对本发明实施例提供的计算机可读存储介质进行介绍,下文描述的计算机可读存储介质与上文描述的图像存储方法或图像读取方法可相互对应参照。
本发明还提供一种计算机可读存储介质,计算机可读存储介质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的图像存储方法的步骤或上述的图像读取方法的步骤。
该计算机可读存储介质可以包括:U盘、移动硬盘、只读存储器(Read-Only Memory,ROM)、随机存取存储器(Random Access Memory,RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
本说明书中各个实施例采用递进的方式描述,每个实施例重点说明的都是与其它实施例的不同之处,各个实施例之间相同或相似部分互相参见即可。对于实施例公开的装置而言,由于其与实施例公开的方法相对应,所以描述的比较简单,相关之处参见方法部分说明即可。
专业人员还可以进一步意识到,结合本文中所公开的实施例描述的各示例的单元及算法步骤,能够以电子硬件、计算机软件或者二者的结合来实现,为了清楚地说明硬件和软件的可互换性,在上述说明中已经按照功能一般性地描述了各示例的组成及步骤。这些功能究竟以硬件还是软件的方式来执行,取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能,但是这种实现不应该认为超出本发明的范围。
结合本文中所公开的实施例描述的方法或算法的步骤可以直接用硬件、处理器执行的软件模块,或者二者的结合来实施。软件模块可以置于随机存储器(RAM)、内存、只读存储器(ROM)、电可编程ROM、电可擦除可编程ROM、寄存器、硬盘、可移动磁盘、CD-ROM、或技术领域内所公知的任意其它形式的存储介质中。
最后,还需要说明的是,在本文中,诸如第一和第二等之类的关系属于仅仅用来将一个实体或者操作与另一个实体或者操作区分开来,而不一定要求或者暗示这些实体或操作之间存在任何这种实际的关系或者顺序。而且,术语“包括”、“包含”或者其他任何变体意在涵盖非排他性的包含,从而使得包括一系列要素的过程、方法、物品或者设备不仅包括那些要素,而且还包括没有明确列出的其他要素,或者是还包括为这种过程、方法、 物品或者设备所固有的要素。
以上对本发明所提供的一种图像存储方法、图像读取方法、图像存储器、图像存储装置、图像读取装置及计算机可读存储介质进行了详细介绍,本文中应用了具体个例对本发明的原理及实施方式进行了阐述,以上实施例的说明只是用于帮助理解本发明的方法及其核心思想;同时,对于本领域的一般技术人员,依据本发明的思想,在具体实施方式及应用范围上均会有改变之处,综上所述,本说明书内容不应理解为对本发明的限制。

Claims (13)

  1. 一种图像存储方法,其特征在于,包括:
    获取目标图像,对所述目标图像进行分解,得到目标子图像;
    将所述目标子图像与元数据存储区中的元数据进行匹配,确定所述目标子图像对应的目标元数据;
    获取所述目标元数据对应的目标元数据信息和所述目标图像的属性信息;
    利用所述属性信息和所述目标元数据信息组成所述目标图像的组合信息,并存储所述组合信息。
  2. 根据权利要求1所述的图像存储方法,其特征在于,所述获取所述目标元数据对应的目标元数据信息,包括:
    获取所述目标子图像与对应的所述目标元数据之间的变换系数;
    利用元数据表获取所述目标元数据在所述元数据存储区中的预存地址,利用所述变换系数和所述预存地址组成所述目标元数据信息。
  3. 根据权利要求2所述的图像存储方法,其特征在于,所述元数据表的建立过程,包括:
    将各个所述元数据存入所述元数据存储区,并获取各个所述元数据对应的所述预存地址;
    利用各个所述预存地址构建所述元数据表。
  4. 根据权利要求3所述的图像存储方法,其特征在于,在所述将所述目标子图像与元数据存储区中的元数据进行匹配之后,还包括:
    将特殊子图像存入所述元数据存储区,并获取所述特殊子图像对应的预存地址;其中,所述特殊子图像为匹配失败的目标子图像;
    利用所述特殊子图像对应的所述预存地址更新所述元数据表。
  5. 根据权利要求4所述的图像存储方法,其特征在于,所述将特殊子图像存入所述元数据存储区,包括:
    对所述特殊子图像进行压缩处理后存入所述元数据存储区。
  6. 根据权利要求1所述的图像存储方法,其特征在于,所述获取所述目标图像的属性信息,包括:
    获取所述目标子图像在所述目标图像中对应的位置信息;
    利用所述位置信息和所述目标图像的标识信息组成所述属性信息。
  7. 一种图像读取方法,其特征在于,包括:
    获取用于读取目标图像的读取指令,利用所述读取指令确定所述目标图像对应的组合信息;
    对所述组合信息进行解析,得到属性信息和目标元数据信息,利用所述目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据;
    利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建处理,得到所述目标图像。
  8. 根据权利要求7所述的图像读取方法,其特征在于,所述利用所述目标元数据信息从元数据存储区中获取对应的目标元数据,包括:
    对所述目标元数据信息进行解析,得到对应的预存地址;
    利用所述预存地址从所述元数据存储区中获取对应的所述目标元数据。
  9. 根据权利要求7所述的图像读取方法,其特征在于,所述利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建处理,得到所述目标图像,包括:
    对所述目标元数据信息进行解析,得到变换系数;
    利用所述变换系数对对应的所述目标元数据进行处理,得到对应的目标子图像;
    对所述属性信息进行解析,得到所述目标子图像对应的位置信息;
    利用所述位置信息对所述目标子图像进行构建处理,得到所述目标图像。
  10. 一种图像存储器,其特征在于,包括处理器、存储器和输入输出部件,其中:
    所述输入输出部件,用于获取或输出目标图像;
    所述存储器,包括元数据存储区、组合信息存储区和程序存储区;其中,所述程序存储区用于保存计算机程序,所述元数据存储区用于保存元 数据,所述组合信息存储区用于存储组合信息;
    所述处理器,用于执行所述计算机程序,以实现如权利要求1至6任一项所述的图像存储方法或如权利要求7至9任一项所述的图像读取方法。
  11. 一种图像存储装置,其特征在于,包括:
    分解模块,用于获取目标图像,对所述目标图像进行分解,得到目标子图像;
    匹配模块,用于将所述目标子图像与元数据存储区中的元数据进行匹配,确定所述目标子图像对应的目标元数据;
    获取模块,用于获取所述目标元数据对应的目标元数据信息和所述目标图像的属性信息;
    存储模块,用于利用所述属性信息和所述目标元数据信息组成所述目标图像的组合信息,并存储所述组合信息。
  12. 一种图像读取装置,其特征在于,包括:
    指令获取模块,用于获取用于读取目标图像的读取指令,利用所述读取指令确定所述目标图像对应的组合信息;
    解析模块,用于对所述组合信息进行解析,得到属性信息和目标元数据信息,利用所述目标元数据信息从元数据存储区中的多个元数据中获取对应的目标元数据;
    目标图像获取模块,用于利用所述属性信息和所述目标元数据信息对所述目标元数据进行构建处理,得到所述目标图像。
  13. 一种计算机可读存储介质,其特征在于,用于保存计算机程序,其中,所述计算机程序被处理器执行时实现如权利要求1至6任一项所述的图像存储方法或如权利要求7至9任一项所述的图像读取方法。
PCT/CN2020/087161 2019-11-07 2020-04-27 图像存储方法、图像读取方法、图像存储器及存储介质 WO2021088333A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201911082402.X 2019-11-07
CN201911082402.XA CN110909187B (zh) 2019-11-07 2019-11-07 图像存储方法、图像读取方法、图像存储器及存储介质

Publications (1)

Publication Number Publication Date
WO2021088333A1 true WO2021088333A1 (zh) 2021-05-14

Family

ID=69816354

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/087161 WO2021088333A1 (zh) 2019-11-07 2020-04-27 图像存储方法、图像读取方法、图像存储器及存储介质

Country Status (2)

Country Link
CN (1) CN110909187B (zh)
WO (1) WO2021088333A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110909187B (zh) * 2019-11-07 2023-01-10 苏州浪潮智能科技有限公司 图像存储方法、图像读取方法、图像存储器及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6912531B1 (en) * 1999-07-30 2005-06-28 Canon Kabushiki Kaisha Image storage method and apparatus
CN108885782A (zh) * 2017-08-09 2018-11-23 深圳市大疆创新科技有限公司 图像处理方法、设备及计算机可读存储介质
CN109543062A (zh) * 2018-09-29 2019-03-29 中国平安人寿保险股份有限公司 图像处理方法、系统、计算机装置及可读存储介质
CN109992200A (zh) * 2017-12-30 2019-07-09 苏龙田 一种使信息存储占用空间最小的方法
CN110909187A (zh) * 2019-11-07 2020-03-24 苏州浪潮智能科技有限公司 图像存储方法、图像读取方法、图像存储器及存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103873438A (zh) * 2012-12-12 2014-06-18 鸿富锦精密工业(深圳)有限公司 压缩包上传去重系统及方法
CN103873504A (zh) * 2012-12-12 2014-06-18 鸿富锦精密工业(深圳)有限公司 数据分块存储至分布式服务器的系统及方法
CN110275861B (zh) * 2019-06-25 2021-09-17 北京明略软件系统有限公司 数据存储方法及装置、存储介质、电子装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6912531B1 (en) * 1999-07-30 2005-06-28 Canon Kabushiki Kaisha Image storage method and apparatus
CN108885782A (zh) * 2017-08-09 2018-11-23 深圳市大疆创新科技有限公司 图像处理方法、设备及计算机可读存储介质
CN109992200A (zh) * 2017-12-30 2019-07-09 苏龙田 一种使信息存储占用空间最小的方法
CN109543062A (zh) * 2018-09-29 2019-03-29 中国平安人寿保险股份有限公司 图像处理方法、系统、计算机装置及可读存储介质
CN110909187A (zh) * 2019-11-07 2020-03-24 苏州浪潮智能科技有限公司 图像存储方法、图像读取方法、图像存储器及存储介质

Also Published As

Publication number Publication date
CN110909187A (zh) 2020-03-24
CN110909187B (zh) 2023-01-10

Similar Documents

Publication Publication Date Title
CN110443140B (zh) 文本定位的方法、装置、计算机设备及存储介质
WO2022028313A1 (en) Method and device for image generation and colorization
CN114005012A (zh) 多模态预训练模型的训练方法、装置、设备及存储介质
US10699751B1 (en) Method, system and device for fitting target object in video frame
WO2017088479A1 (zh) 台标识别方法及装置
KR20200059993A (ko) 웹툰 제작을 위한 콘티를 생성하는 장치 및 방법
CN113781356B (zh) 图像去噪模型的训练方法、图像去噪方法、装置及设备
WO2019041442A1 (zh) 图表数据结构化提取方法、系统、电子设备及计算机可读存储介质
CN111918065A (zh) 一种信息压缩/解压缩方法及装置
CN110662080B (zh) 面向机器的通用编码方法
CN113343958A (zh) 一种文本识别方法、装置、设备及介质
WO2021088333A1 (zh) 图像存储方法、图像读取方法、图像存储器及存储介质
CN111767889A (zh) 公式识别方法、电子设备及计算机可读介质
CN110012338A (zh) 一种显示弹幕数据的方法、装置、计算机设备和存储介质
CN110717060B (zh) 图像mask的过滤方法、装置及存储介质
WO2023142438A1 (zh) 终端的碎屏识别方法和相关设备
WO2020155299A1 (zh) 视频帧中目标对象的拟合方法、系统及设备
JP2007109177A (ja) 画像処理装置及びその制御方法、プログラム
JP2012060452A (ja) 画像処理装置、その方法およびプログラム
JP5159588B2 (ja) 画像処理装置、画像処理方法、コンピュータプログラム
CN115273057A (zh) 文本识别方法、装置和听写批改方法、装置及电子设备
CN110996026B (zh) 一种osd显示方法、装置、设备及存储介质
CN115843375A (zh) 徽标标注方法及装置、徽标检测模型更新方法及系统、存储介质
CN110909726A (zh) 一种基于图像识别的书面文档互动系统和方法
CN116798056B (zh) 表格图像定位方法、装置、设备、计算机可读存储介质

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: 20884359

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20884359

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 20884359

Country of ref document: EP

Kind code of ref document: A1