CN107292934A - Image content recognizing method, modeling method and relevant apparatus - Google Patents
Image content recognizing method, modeling method and relevant apparatus Download PDFInfo
- Publication number
- CN107292934A CN107292934A CN201710448132.4A CN201710448132A CN107292934A CN 107292934 A CN107292934 A CN 107292934A CN 201710448132 A CN201710448132 A CN 201710448132A CN 107292934 A CN107292934 A CN 107292934A
- Authority
- CN
- China
- Prior art keywords
- picture
- color
- color distribution
- content information
- identified
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 103
- 238000009826 distribution Methods 0.000 claims abstract description 239
- 238000012545 processing Methods 0.000 claims abstract description 25
- 239000000463 material Substances 0.000 claims description 55
- 238000004458 analytical method Methods 0.000 claims description 27
- 238000010801 machine learning Methods 0.000 claims description 16
- 239000003086 colorant Substances 0.000 description 23
- 238000005457 optimization Methods 0.000 description 14
- 238000010586 diagram Methods 0.000 description 9
- 230000001788 irregular Effects 0.000 description 5
- 241000227653 Lycopersicon Species 0.000 description 4
- 235000007688 Lycopersicon esculentum Nutrition 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 241000972773 Aulopiformes Species 0.000 description 2
- 241000657513 Senna surattensis Species 0.000 description 2
- 235000013601 eggs Nutrition 0.000 description 2
- 235000019515 salmon Nutrition 0.000 description 2
- 241000251468 Actinopterygii Species 0.000 description 1
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000021186 dishes Nutrition 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 235000019688 fish Nutrition 0.000 description 1
- 239000000796 flavoring agent Substances 0.000 description 1
- 235000019634 flavors Nutrition 0.000 description 1
- 235000012055 fruits and vegetables Nutrition 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 235000015277 pork Nutrition 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- 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
- G06T2207/10024—Color image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the present invention provides a kind of image content recognizing method, modeling method and relevant apparatus, is related to picture processing field and data processing field.Wherein, the image content recognizing method includes：Analyze picture to be identified and determine its COLOR COMPOSITION THROUGH DISTRIBUTION；The COLOR COMPOSITION THROUGH DISTRIBUTION and reference color for contrasting the picture to be identified are distributed, it is determined that first reference color corresponding with the COLOR COMPOSITION THROUGH DISTRIBUTION of the picture to be identified is distributed；First reference color is distributed to corresponding content information as the image content of the picture to be identified.Technical scheme provided in an embodiment of the present invention can recognize image content.
Description
Technical Field
The embodiment of the invention relates to the field of picture processing and data processing, in particular to a picture content identification method, a modeling method and a related device.
Background
In the prior art, a method for recognizing pictures is mostly adopted in the aspect of picture content recognition. That is, after a picture is uploaded, some pictures close to the pictures are returned for the uploader to refer to determine the content of the picture. The method can not directly determine the picture content, and has large data retrieval amount and high cost.
Disclosure of Invention
The embodiment of the invention provides a picture content identification method, a modeling method and a related device, which are used for solving the problem that the picture content cannot be effectively identified in the prior art.
In a first aspect, an embodiment of the present invention provides a picture content identification method, including:
analyzing the picture to be identified to determine the color distribution of the picture;
comparing the color distribution of the picture to be identified with a reference color distribution, and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
In a second aspect, an embodiment of the present invention provides an apparatus for identifying picture content, including:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the reference determining module is used for comparing the color distribution of the picture to be identified with a reference color distribution and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and the content determining module is used for taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
In a third aspect, an embodiment of the present invention provides a modeling method, including:
acquiring a sample picture of known content information;
analyzing the sample picture to determine the color distribution thereof;
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
In a fourth aspect, an embodiment of the present invention provides a machine learning apparatus, including:
the acquisition module is used for acquiring a sample picture of known content information;
the analysis module is used for analyzing the sample picture to determine the color distribution of the sample picture;
a processing module to:
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
In a fifth aspect, an embodiment of the present invention provides a method for identifying picture content, including:
analyzing the picture to be identified to determine the color distribution of the picture;
querying a material library to determine a combination of materials corresponding to the color distribution of the picture to be identified;
determining content information of the picture to be identified based on the combination of the materials.
In a sixth aspect, an embodiment of the present invention provides an apparatus for identifying picture content, including:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the combination determining module is used for inquiring a material library to determine the combination of materials corresponding to the color distribution of the picture to be identified;
a content determination module for determining content information of the picture to be identified based on the combination of the materials.
In a seventh aspect, an embodiment of the present invention provides an apparatus for identifying picture content, including a memory and a processor; wherein the memory is configured to store one or more computer instructions; wherein the one or more computer instructions are for execution invoked by the processor; wherein the processor is configured to execute the instructions to implement the method according to the first aspect of the embodiment of the present invention.
In an eighth aspect, an embodiment of the present invention provides a machine learning apparatus, including a memory and a processor; wherein the memory is configured to store one or more computer instructions; wherein the one or more computer instructions are for execution invoked by the processor; wherein the processor is configured to execute the instructions to implement the method according to the third aspect of the embodiment of the present invention.
In a ninth aspect, an embodiment of the present invention provides an apparatus for identifying picture content, including a memory and a processor; wherein the memory is configured to store one or more computer instructions; wherein the one or more computer instructions are for execution invoked by the processor; wherein the processor is configured to execute the instructions to implement the method according to the fifth aspect of the embodiment of the present invention.
In a tenth aspect, embodiments of the present invention provide a computer storage medium storing one or more computer instructions, wherein the instructions, when executed, implement a method according to the first, third or fifth aspect of embodiments of the present invention.
The various embodiments of the invention can identify the picture content or provide a basis for effectively and accurately identifying the picture content.
These and other aspects of the invention are apparent from and will be elucidated with reference to the embodiments described hereinafter.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 shows a flow chart of a picture content identification method according to an embodiment of the invention;
FIG. 2 is a flow chart illustrating a method for determining a color distribution of a picture to be recognized according to an embodiment of the present invention;
FIG. 3 shows a block diagram of a picture content recognition apparatus according to an embodiment of the present invention;
FIG. 4 is a block diagram of a color analysis module of the picture content recognition apparatus shown in FIG. 3;
FIG. 5 shows a flow diagram of a modeling method according to an embodiment of the invention;
FIG. 6 shows a block diagram of a machine learning apparatus according to one embodiment of the invention;
FIG. 7 illustrates a block diagram of a color analysis module of the machine learning device of FIG. 6;
FIG. 8 is a flow chart illustrating a picture content identification method according to another embodiment of the present invention;
fig. 9 shows a block diagram of a picture content recognition apparatus according to another embodiment of the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention.
In some of the flows described in the present specification and claims and in the above figures, a number of operations are included that occur in a particular order, but it should be clearly understood that these operations may be performed out of order or in parallel as they occur herein, with the order of the operations being indicated as 101, 102, etc. merely to distinguish between the various operations, and the order of the operations by themselves does not represent any order of performance. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor limit the types of "first" and "second" to be different.
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. The embodiments of the present invention and the embodiments obtained based on the idea of the present invention are within the scope of protection of the present invention.
Fig. 1 is a flowchart illustrating a method for identifying picture content according to an embodiment of the present invention. The picture content refers to the content actually expressed by the picture, for example, the picture of the fruit and vegetable, and the content is information such as a specific fruit name; the contents of the dish picture are specific names of dishes and the like. Referring to fig. 1, the method includes:
102: and analyzing the picture to be identified to determine the color distribution of the picture.
Optionally, in an implementation manner of this embodiment, the to-be-recognized picture is obtained by using the prior art. For example, by cell phone scanning, AR (Augmented Reality) device scanning, etc.
104: and comparing the color distribution of the picture to be identified with a reference color distribution, and determining a first reference color distribution corresponding to the color distribution of the picture to be identified.
In the present embodiment, the reference color distributions are predetermined, and each of the reference color distributions corresponds to one type of content information.
Optionally, in an implementation manner of this embodiment, the first reference color distribution corresponds to a color distribution of the picture to be recognized, and includes the following scheme: the first reference color distribution is the same as the color distribution of the picture to be recognized, the first reference color distribution is similar to the color distribution of the picture to be recognized (for example, satisfies a set deviation value, or is determined to be similar by similarity calculation), or the color distribution of the picture to be recognized falls within the range of the first reference color distribution.
106: and taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
By adopting the method provided by the embodiment, the picture content can be effectively identified, and the quality (e.g. accuracy) of picture content identification is improved. For example, the method is applied to the field of take-out, and the name of the dish expressed by the dish picture can be accurately identified.
Optionally, in an implementation manner of this embodiment, as shown in a dashed box in fig. 1, before the process 102, the method further includes a process 100: and optimizing the source picture to obtain the picture to be identified with the reserved effective area. Wherein, the effective area refers to an area containing picture content.
For example, pictures in many areas, such as take-away, have size requirements and position specifications for the picture content. Based on these specifications, the effective area can be reserved by cropping the picture to a specified size. For another example, background colors of pictures in many fields have no influence on the recognition of the contents of the pictures, and therefore, the background color portion may be deleted in the process 100. In summary, the purpose of process 100 is to preserve the active area and thereby reduce the data throughput of subsequent processing. Based on the purpose, different optimization modes can be designed by combining the characteristics of pictures in the application field, and the optimization modes are not listed in detail in the invention.
By adopting the implementation mode, invalid calculation caused by invalid areas during picture processing can be reduced, and the picture processing efficiency is improved.
Fig. 2 is a flowchart illustrating a method for determining a color distribution of a picture to be recognized according to an embodiment of the present invention. Referring to fig. 2, the method includes:
200: and identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value.
Optionally, in a first implementation manner of this embodiment, the coordinate system may be a commonly-used planar rectangular coordinate system or a custom coordinate system. The color value-to-color correspondence (i.e., color standard) may conform to a commonly used color standard (e.g., an existing R (red) G (green) B (blue) color standard) or a custom color standard. The identified location may be a point or a region (e.g., a square region) in a coordinate system, etc.
Alternatively, in a second implementation of the embodiment, a color range is set for each color, for example, rgb (200, 10) represents the color value of yellow, and then rgb fluctuation 20 can be selected as a threshold value of the yellow range. And the rest colors are analogized, and finally a color threshold value table is formed. In this implementation manner, which colors can be customized is specifically set, for example, a specified color is set according to the characteristics of the picture in the application field, and a commonly used color may also be set.
202: and determining the color distribution of the picture to be recognized according to the color values of the picture to be recognized at each recognition position. The color distribution includes different color blocks, each color block includes: color, shape, and relative position to other patches (optional).
Specifically, by determining the color value of each recognition position, the color block included in the picture to be recognized can be determined according to the color value and the recognition position. For example, in the case that the process 200 adopts the first implementation manner, in this embodiment, the colors at the respective recognition positions may be classified based on the set color value ranges of the colors, and color patches may be finally formed. In the case where the second implementation is adopted in the process 200, the color value of the recognition position is determined, and at the same time, the color patch to which the recognition position belongs is also determined.
In process 202, the color block contains the following attributes: color and shape. Furthermore, a color block may also contain the following properties: a position in a picture, or a relative position to other or specified tiles in the same picture (e.g., the tile with the largest area, the tile with the center position, etc.). Wherein, the relative position may contain any one or more of the following information: the orientation relationship between color patches, the surrounding relationship, the semi-surrounding relationship, the spacing (with other color patches in between) relationship, the contact relationship, etc.
By adopting the embodiment, the color distribution of the picture to be identified can be effectively determined. Generally, the quality of the determined color distribution directly affects the quality of picture content identification, and a certain amount of data calculation is required as cost to obtain a high-quality color distribution. Therefore, one skilled in the art can obtain color distributions of different qualities as desired. For example, only the color and the shape of the patch are determined without determining the relative positional relationship of the patch, or the color, the shape, the relative positional relationship, and the like of the patch are determined at the same time.
Optionally, in an implementation manner of this embodiment, the color distribution optimization may be performed in the following manner: and deleting the color blocks of which the area ratio does not meet the preset condition, or replacing the color blocks with the colors of similar color blocks. For example, if the area ratio of a certain color patch is smaller than a set threshold value, or the area ratio is smaller than a set value compared with other color patches, the color patch is deleted (that is, the color patch is not used as a part of the color distribution), or the color patch is replaced with the color of a similar color patch (for example, a color patch adjacent to the color patch and having the longest contact boundary), and accordingly, the shape of the subsequently determined similar color patch is relatively changed.
Optionally, in an implementation manner of this embodiment, the color distribution optimization may be performed in the following manner: and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block. For example, if a color block is located only at a corner of a picture, the color block is deleted, or the color block is replaced with the color of a similar color block (e.g., a color block adjacent to the color block and having the longest contact boundary).
Optionally, in an implementation manner of this embodiment, the color distribution optimization may be performed by combining the foregoing two implementation manners. For example, color blocks which are only located at corners and have an area ratio smaller than a preset threshold are deleted or replaced by colors of similar color blocks; and deleting or replacing the color blocks which are contained by the large-area color blocks and have the areas smaller than a set value or a set proportion with the colors of the similar color blocks.
By optimizing the color distribution, noise interference is avoided, and the identification quality of the picture content is improved.
In an embodiment of the present invention, the method for determining the color distribution of the picture to be recognized shown in fig. 2 may be implemented as a specific implementation of the process 102 in the picture content recognition method shown in fig. 1. In this embodiment, each of the reference color distributions corresponds to one type of content information, and each of the reference color distributions includes a color value range, a shape set, and a relative position set (optional).
In a specific implementation manner of this embodiment, the reference color distribution may include at least two color value ranges, and each color value range corresponds to the shape set and the relative position set (which are optional). Each color value range may be defined separately, may be the same as the color value range of the color patches set in the processes 200 and 202, and may be larger than the color value range of the color patches set in the processes 200 and 202 by a reasonable range. This can be selected or improved by one skilled in the art as desired.
Optionally, in an implementation manner of this embodiment, the reference color distribution includes the following contents: color patch a (color value range a0-a1) -shape (oval, circle, quadrilateral) -relative position (contained in, semi-surrounded by color patch B); color patch B (color value range B0-B1) -shape (irregular polygon) -relative position (containing color patch a, half surrounding color patch a, located in the three o' clock direction of color patch C) -color patch C … ….
In this embodiment, the processing 106 may specifically be implemented in the following manner:
determining the reference color distribution which has the same number of color blocks as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution is used as the first reference color distribution, and the color value range and the shape set both cover (i.e., contain or are equal to) the colors and the shapes of the color blocks in the to-be-recognized picture; or, determining the reference color distribution with the same number of color blocks as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition search, wherein the reference color distribution with the color value range, the shape set and the relative position set covering the colors, the shapes and the relative positions of the color blocks in the to-be-recognized picture is used as the first reference color distribution.
By adopting the embodiment, the picture content can be effectively identified, and the quality (e.g. accuracy) of picture content identification is improved.
In one embodiment of the present invention, a picture content recognition apparatus is provided that includes a memory and a processor. Wherein the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke execution; the processor is configured to execute the instructions to implement the picture content identification method provided by the foregoing embodiments or implementations.
Fig. 3 is an example of a block diagram of a picture content recognition apparatus according to an embodiment of the present invention. Referring to fig. 3, the picture content recognition apparatus includes a color analysis module 32, a reference determination module 34, and a content determination module 36. The details will be described below.
In the present embodiment, the color analysis module 32 is configured to analyze the picture to be recognized to determine the color distribution thereof. The reference determining module 34 is configured to compare the color distribution of the to-be-identified picture with a reference color distribution, and determine a first reference color distribution corresponding to the color distribution of the to-be-identified picture. The content determining module 36 is configured to use content information corresponding to the first reference color distribution as picture content of the picture to be identified.
By adopting the picture content identification device provided by the embodiment, the picture content can be effectively identified, and the quality (e.g. accuracy) of picture content identification is improved. For example, the method is applied to the field of take-out, and the name of the dish expressed by the dish picture can be accurately identified.
Optionally, in an implementation manner of this embodiment, as shown by a dashed box in fig. 3, the picture content identifying apparatus further includes a picture optimizing module 30, configured to optimize a source picture to obtain the picture to be identified with an effective area reserved. By adopting the implementation mode, the subsequent data processing amount can be reduced, and the picture identification efficiency is improved.
Optionally, in an implementation manner of this embodiment, as shown in fig. 4, the color analysis module 32 includes: and the identifying submodule 322 is configured to identify the color of the to-be-identified picture at each identification position based on the coordinate system and the color value. A determination sub-module 324 for: determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position; or determining the color and the shape of the color blocks in the picture to be recognized and the relative positions of part or all of the color blocks according to the color values of the picture to be recognized at the recognition positions.
Optionally, as shown by a dashed line box in fig. 4, the color analysis module 32 further includes an optimization submodule 326, configured to delete a color block whose area ratio does not satisfy a preset condition, or replace the color block with a color of a similar color block; and/or deleting the color blocks of which the positions do not meet the preset conditions, or replacing the color blocks with the colors of the similar color blocks.
Optionally, in an implementation manner of this embodiment, each of the reference color distributions corresponds to one type of content information. Each of the reference color distributions includes a range of color values and a set of shapes, or each of the reference color distributions includes a range of color values, a set of shapes, and a set of relative positions.
Optionally, in an implementation manner of this embodiment, the content determining module is specifically configured to:
determining the reference color distribution which has the same number of color blocks as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution is used as the first reference color distribution, and the color value range and the shape set both cover (i.e., contain or are equal to) the colors and the shapes of the color blocks in the to-be-recognized picture; or, determining the reference color distribution with the same number of color blocks as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition search, wherein the reference color distribution with the color value range, the shape set and the relative position set covering the colors, the shapes and the relative positions of the color blocks in the to-be-recognized picture is used as the first reference color distribution.
In this embodiment, for the explanation of the related nouns, ranges, etc., the detailed description of the processes executed or executable by each module and sub-module, and the detailed description of the detailed logic processing procedure, please refer to the detailed description in the picture content identification method provided in various embodiments of the present invention, which is not repeated herein.
Fig. 5 is a schematic flow chart of a modeling method (which can also be understood as a data processing method) according to an embodiment of the present invention, and referring to fig. 5, the method includes:
500: a sample picture of known content information is obtained.
Optionally, in an implementation manner of this embodiment, a plurality of sample pictures corresponding to each type of content information are input into the machine learning device.
504: analyzing the sample picture to determine its color distribution.
Optionally, in an implementation manner of this embodiment, the color distribution may include: the pictures contain color blocks, the shapes of the color blocks and the relative positions of part or all of the color blocks (optional).
506: and determining or updating a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
The corresponding relationship between the content information determined/updated by the method provided by the embodiment and the reference color distribution can be used as a model or template, so that the picture content can be identified based on the model or template.
Optionally, in an implementation manner of this embodiment, as shown in a dashed box in fig. 5, before analyzing the sample picture, the method further includes processing 502: optimizing the sample picture to preserve an active area of the sample picture. The implementation mode is beneficial to reducing unnecessary data operation.
Optionally, in an implementation manner of this embodiment, the process 504 is implemented in the following manner: identifying the color of the sample picture at each identification position based on the coordinate system and the color value; and determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or determining the color and the shape of the color block in the picture to be recognized and the relative position of part or all of the color blocks according to the color value of the picture to be recognized at each recognition position.
In this implementation, the following processing may also be performed: deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or deleting the color blocks of which the positions do not meet the preset conditions, or replacing the color blocks with the colors of the similar color blocks.
Optionally, in an implementation manner of this embodiment, the processing 506 includes: and statistically determining the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture. Illustratively, this may be achieved in the following way: counting the color distribution of a plurality of sample pictures under the same content information; and taking the color value range and the shape set which are determined by statistics as the reference color distribution of the same content information, or taking the color value range, the shape set and the relative position set which are determined by statistics as the reference color distribution of the same content information.
For example, assume that, in two (for example only, generally, the more the number of pictures is, the more accurate the result is) pictures corresponding to the dish name of "shredded pork with fish flavor"), the picture P1 includes color block 1 (color value range is M) and color block 2 (color value range is N), the shape of color block 1 is circular, and the shape of color block 2 is irregular polygon and surrounds color block 1; the picture P2 includes the color block 1 and the color block 2, where the shape of the color block 1 is an irregular polygon, and the shape of the color block 2 is an ellipse and surrounds the color block 1. Then, from these two pictures, the reference color distribution can be counted as: color block 1 (color value range M) -shape (circle, irregular polygon) -relative position (enclosed by color block 2); color block 2 (color value range N) -shape (ellipse, irregular polygon) -relative position (surrounding color block 1).
Of course, the above is only a simple example, and actually, due to the large number of pictures, the statistics can be performed by using the existing statistical principle. The present invention is not illustrated in detail.
Optionally, in an implementation manner of this embodiment, the processing 506 includes: and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture. Illustratively, this may be achieved in the following way: comparing the color distribution of the sample picture with the reference color distribution; if there is a difference attribute in the color of the sample picture with respect to the reference color distribution (i.e., an attribute that the sample picture has but not the reference color distribution), the difference attribute is added to the reference color distribution. Wherein the color distribution of the sample picture comprises the following attributes: the color and shape of the color blocks, or the relative positions of part or all of the color blocks. Wherein the reference color distribution includes the following attributes: a range of color values and a set of shapes, or a set of relative positions.
In the present embodiment, for the explanation of the related terms, ranges, applications, etc., please refer to the related description in the picture content method, which is not repeated herein.
Fig. 6 is a block diagram of a machine learning apparatus according to an embodiment of the present invention. Referring to fig. 6, the machine learning apparatus includes an acquisition module 60, an analysis module 64, and a processing module 66. The details will be described below.
In this embodiment, the obtaining module 60 is configured to obtain a sample picture of known content information. The analysis module 64 is used for analyzing the sample picture to determine its color distribution. The processing module 66 is configured to: based on the content information of the sample picture and the color distribution of the sample picture, counting and determining a reference color distribution corresponding to the content information; or updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
With the machine learning apparatus provided in this embodiment, the correspondence between the content information and the reference color distribution can be determined/updated, so as to perform picture content identification according to the correspondence (which can be understood as a template or a model).
Optionally, in an implementation manner of this embodiment, as shown in a dashed box in fig. 6, the machine learning apparatus further includes an optimizing module 62, configured to optimize the sample picture to reserve an effective area of the sample picture before the analyzing module analyzes the sample picture.
Optionally, in an implementation manner of this embodiment, as shown in fig. 7, the analysis module 64 includes: the identification submodule 642 is used for identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value; the determining submodule 644 is configured to determine, according to color values of the to-be-identified picture at the respective identification positions, colors and shapes of color blocks in the to-be-identified picture; or determining the color and the shape of the color blocks in the picture to be recognized and the relative positions of part or all of the color blocks according to the color values of the picture to be recognized at the recognition positions.
Optionally, as shown by a dashed box in fig. 7, the analysis module 64 further includes an optimization sub-module 646, configured to delete a color block whose area ratio does not satisfy a preset condition, or replace the color block with a color of a similar color block; and/or deleting the color blocks of which the positions do not meet the preset conditions, or replacing the color blocks with the colors of the similar color blocks.
Optionally, in an implementation manner of this embodiment, the processing module 66 includes a reference determination sub-module, configured to perform the following processing: counting the color distribution of a plurality of sample pictures under the same content information; and using the color value range and the shape set determined by statistics as the reference color distribution of the same content information, or using the color value range, the shape set and the relative position set determined by statistics as the reference color distribution of the same content information.
Optionally, in a practice manner of this embodiment, the processing module includes a reference updating sub-module, configured to compare and determine a difference attribute of the color distribution of the sample picture with respect to the reference color distribution, and add the difference attribute to the reference color distribution. The color distribution of the sample picture comprises the colors and the shapes of the color blocks, or also comprises the relative positions of part or all of the color blocks. Accordingly, the reference color distribution includes a range of color values and a set of shapes, or alternatively, also includes a set of relative positions.
In the embodiment, for the explanations of the terms, ranges, applications, and the like in the machine learning device, the details of the processes executed or executable by each module and sub-module are please refer to the related descriptions in the modeling method provided above, which are not repeated herein.
In one embodiment of the present invention, a machine learning apparatus is provided that includes a memory and a processor. Wherein the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke execution; the processor is used for executing the instructions to realize the modeling method provided by the embodiment or the implementation mode as described in the foregoing.
Fig. 8 is a flowchart illustrating a method for identifying picture content according to an embodiment of the present invention, and referring to fig. 8, the method includes:
802: and analyzing the picture to be identified to determine the color distribution of the picture.
For the description of the processing 802, please refer to the related description above, which is not repeated herein.
Optionally, before 802, optimizing a source picture to obtain the picture to be identified with a reserved effective area.
804: and querying a material library to determine the combination of materials corresponding to the color distribution of the picture to be identified.
In the present embodiment, "material" refers to the composition of picture content. Taking the picture with the picture content of salmon sushi as an example, the material can be rice and salmon.
Optionally, in an implementation manner of this embodiment, the material library includes a color value range and a shape set corresponding to each of the plurality of materials; or, the materials library includes a range of color values, a set of shapes, and relative positions for each of a plurality of materials. The relative position here refers to the relative position between the materials in a fixed arrangement of the materials. Accordingly, when the relative position is included in the color distribution of the picture to be recognized, the material possibly included in the picture to be recognized can be further determined.
806: determining content information of the picture to be identified based on the combination of the materials.
Optionally, in an implementation manner of this embodiment, the determined material names are combined according to a common material combination relationship to obtain content information of the picture. Wherein the material combination relationship may include: the materials may be combined, combinations of materials (e.g., fried, stir-fried, steamed, fried, etc.).
For example, with a picture whose picture content is tomato scrambled eggs, when 804 is used to determine that the contained material includes tomatoes and eggs, at 806 tomato scrambled eggs are returned as picture content according to the usual combination of tomatoes and eggs.
Of course, when there are a plurality of material combination relationships, a plurality of picture contents may be output as the recognition result in 806. For example, the preferred picture content and the alternative picture content are output according to the probability of the material combination.
By adopting the method provided by the embodiment, on one hand, the method can be used for identifying the picture content; on the other hand, the method can be applied to machine learning or model training related to picture content recognition.
Fig. 9 is a block diagram of a picture content recognition apparatus according to an embodiment of the present invention, and referring to fig. 9, the picture content recognition apparatus includes: the color analysis module 92 is used for analyzing the picture to be identified to determine the color distribution of the picture; a combination determination module 94, configured to query a material library to determine a combination of materials corresponding to the color distribution of the picture to be recognized; a content determining module 96, configured to determine content information of the picture to be identified based on the combination of the materials.
Optionally, in an implementation manner of this embodiment, as shown by a dashed box in fig. 9, the apparatus further includes a picture optimization module 90, configured to optimize a source picture to obtain the picture to be identified with an effective area reserved.
Optionally, in an implementation manner of this embodiment, the material library includes a color value range and a shape set corresponding to each of the plurality of materials; or, the materials library includes a range of color values, a set of shapes, and relative positions for each of a plurality of materials.
For a detailed description of the processes executed or executable by each module and sub-module in the image content recognition apparatus provided in this embodiment, please refer to the description in the embodiment of the method shown in fig. 8, which is not repeated herein.
In one possible design of this embodiment, the picture content recognition device includes a memory and a processor. Wherein the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke execution; the processor is used for executing the instructions to realize the method provided by the embodiment shown in fig. 8 or the implementation manner thereof.
The embodiment of the present invention further provides a computer storage medium, which is used for computer software instructions used by the picture content recognition apparatus/machine learning apparatus provided in the embodiments or the implementation manners of the present invention to implement the functions thereof, and when the instructions are executed, the instructions implement the picture content recognition method or the modeling method provided in the embodiments or the implementation manners of the present invention.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
The invention discloses a1 and a picture content identification method, which comprises the following steps:
analyzing the picture to be identified to determine the color distribution of the picture;
comparing the color distribution of the picture to be identified with a reference color distribution, and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
A2, the method of a1, the method further comprising:
and optimizing the source picture to obtain the picture to be identified with the reserved effective area.
A3, the method as in A1 or A2, the analyzing a picture to be recognized to determine its color distribution, comprising:
identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
A4, the method as in A3, the analyzing the picture to be recognized to determine its color distribution, further comprising:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
A5, method according to A3,
each reference color distribution corresponds to one content information;
each of the reference color distributions includes a range of color values and a set of shapes, or,
each of the reference color distributions includes a range of color values, a set of shapes, and a set of relative positions.
A6, the method as in A5, the comparing the color distribution of the picture to be recognized with a reference color distribution, determining a first reference color distribution corresponding to the color distribution of the picture to be recognized, comprising:
determining color blocks with the same number as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution with a color value range and a shape set covering the colors and shapes of the color blocks in the to-be-recognized picture is used as the first reference color distribution; or,
and determining the reference color distribution which has the same number of color blocks as the to-be-identified picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution is used as the first reference color distribution, and the color value range, the shape set and the relative position set cover the colors, the shapes and the relative positions of the color blocks in the to-be-identified picture.
The invention also discloses B7, a picture content recognition device, the device includes:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the reference determining module is used for comparing the color distribution of the picture to be identified with a reference color distribution and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and the content determining module is used for taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
B8, the apparatus of B7, further comprising:
and the picture optimization module is used for optimizing the source picture to obtain the picture to be identified with the reserved effective area.
B9, the device of B7 or B8, the color analysis module comprising:
the identification submodule is used for identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
a determination submodule for:
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
B10, the apparatus of B9, the color analysis module further comprising an optimization submodule for:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
B11, device according to B9,
each reference color distribution corresponds to one content information;
each of the reference color distributions includes a range of color values and a set of shapes, or,
each of the reference color distributions includes a range of color values, a set of shapes, and a set of relative positions.
B12, the apparatus of B11, the content determining module being configured to:
determining color blocks with the same number as the to-be-recognized picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution with a color value range and a shape set covering the colors and shapes of the color blocks in the to-be-recognized picture is used as the first reference color distribution; or,
and determining the reference color distribution which has the same number of color blocks as the to-be-identified picture from the reference color distribution by means of one-by-one comparison or condition searching, wherein the reference color distribution is used as the first reference color distribution, and the color value range, the shape set and the relative position set cover the colors, the shapes and the relative positions of the color blocks in the to-be-identified picture.
The invention also discloses C13 and a modeling method, wherein the method comprises the following steps:
acquiring a sample picture of known content information;
analyzing the sample picture to determine the color distribution thereof;
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
C14, the method of C13, the method further comprising:
optimizing the sample picture to preserve an active area of the sample picture.
C15, the method of C13, the analyzing the sample picture to determine its color distribution, comprising:
identifying the color of the sample picture at each identification position based on the coordinate system and the color value;
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
C16, the method of C15, the analyzing the sample picture to determine its color distribution, further comprising:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
C17, the method according to C15 or C16, wherein the statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture comprises:
counting the color distribution of a plurality of sample pictures under the same content information;
the statistically determined color value range and shape set are used as the reference color distribution of the same content information, or,
and taking the color value range, the shape set and the relative position set which are determined by statistics as the reference color distribution of the same content information.
C18, the method according to C15 or C16, wherein the updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture comprises:
comparing and determining the difference attribute of the color distribution of the sample picture relative to the reference color distribution;
adding the difference attribute to the reference color distribution;
the color distribution of the sample picture comprises the colors and the shapes of the color blocks, or also comprises the relative positions of part or all of the color blocks;
wherein the reference color distribution includes a range of color values and a set of shapes, or further includes a set of relative positions.
The invention also discloses D19, a machine learning device, comprising:
the acquisition module is used for acquiring a sample picture of known content information;
the analysis module is used for analyzing the sample picture to determine the color distribution of the sample picture;
a processing module to:
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
D20, the apparatus of D19, further comprising:
and the optimization module is used for optimizing the sample picture to reserve the effective area of the sample picture before the analysis module analyzes the sample picture.
D21, the apparatus of D19, the analysis module comprising:
the identification submodule is used for identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
a determination submodule for:
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
D22, the apparatus of D21, the analysis module further comprising an optimization submodule for:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
D23, the apparatus of D21 or D22, the processing module comprising a reference determination submodule to:
counting the color distribution of a plurality of sample pictures under the same content information;
the statistically determined color value range and shape set are used as the reference color distribution of the same content information, or,
and taking the color value range, the shape set and the relative position set which are determined by statistics as the reference color distribution of the same content information.
D24, the apparatus of D21 or D22, the processing module comprising:
a reference updating submodule for determining a difference attribute of the color distribution of the sample picture with respect to the reference color distribution by contrast, and adding the difference attribute to the reference color distribution;
the color distribution of the sample picture comprises the colors and the shapes of the color blocks, or also comprises the relative positions of part or all of the color blocks;
wherein the reference color distribution includes a range of color values and a set of shapes, or further includes a set of relative positions.
The invention also discloses E25 and a picture content identification method, wherein the method comprises the following steps:
analyzing the picture to be identified to determine the color distribution of the picture;
querying a material library to determine a combination of materials corresponding to the color distribution of the picture to be identified;
determining content information of the picture to be identified based on the combination of the materials.
E26, the method of E25, further comprising:
and optimizing the source picture to obtain the picture to be identified with the reserved effective area.
E27, the method as described in E25 or E26, the analyzing a picture to be recognized to determine its color distribution, comprising:
identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
E28, the method of E27, the analyzing a picture to be recognized to determine its color distribution, comprising:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
E29, method as described in E27,
the material library comprises a color value range and a shape set corresponding to each material in a plurality of materials; or,
the materials library includes a range of color values, a set of shapes, and relative positions for each of a plurality of materials.
The invention also discloses F30, a picture content recognition device, the device includes:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the combination determining module is used for inquiring a material library to determine the combination of materials corresponding to the color distribution of the picture to be identified;
a content determination module for determining content information of the picture to be identified based on the combination of the materials.
F31, the apparatus of F30, the apparatus further comprising:
and the picture optimization module is used for optimizing the source picture to obtain the picture to be identified with the reserved effective area.
F32, the device as described in F30 or F31, the color analysis module comprising:
the identification submodule is used for identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
a determination submodule for:
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
F33, the apparatus of F32, the color analysis module further comprising an optimization submodule for:
deleting the color block of which the area ratio does not meet the preset condition, or replacing the color block with the color of a similar color block; and/or the presence of a gas in the gas,
and deleting the color block of which the position does not meet the preset condition, or replacing the color block with the color of the similar color block.
F34, device according to F32,
the material library comprises a color value range and a shape set corresponding to each material in a plurality of materials; or,
the materials library includes a range of color values, a set of shapes, and relative positions for each of a plurality of materials.
The invention also discloses G35, a picture content recognition device, which comprises a memory and a processor; wherein,
the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke for execution;
the processor is configured to execute the instructions to implement the method of any one of claims 1-6.
The invention also discloses H36, a machine learning device, comprising a memory and a processor; wherein,
the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke for execution;
the processor is configured to execute the instructions to implement the method of any one of claims 13-18.
The invention also discloses I37, a picture content recognition device, comprising a memory and a processor; wherein,
the memory is to store one or more computer instructions, wherein the one or more computer instructions are for the processor to invoke for execution;
the processor is configured to execute the instructions to implement the method of any of claims 25-29.
The present invention also discloses J38, a computer storage medium storing one or more computer instructions, wherein the instructions when executed implement:
the method of any one of a1-a 6; or,
the method of any one of C13-C18; or,
the method of any one of E25-E29.
Claims (10)
1. A picture content identification method is characterized by comprising the following steps:
analyzing the picture to be identified to determine the color distribution of the picture;
comparing the color distribution of the picture to be identified with a reference color distribution, and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
2. The method of claim 1, wherein the analyzing the picture to be recognized to determine its color distribution comprises:
identifying the color of the picture to be identified at each identification position based on the coordinate system and the color value;
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
3. The method of claim 2,
each reference color distribution corresponds to one content information;
each of the reference color distributions includes a range of color values and a set of shapes, or,
each of the reference color distributions includes a range of color values, a set of shapes, and a set of relative positions.
4. A picture content recognition device, the device comprising:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the reference determining module is used for comparing the color distribution of the picture to be identified with a reference color distribution and determining a first reference color distribution corresponding to the color distribution of the picture to be identified;
and the content determining module is used for taking the content information corresponding to the first reference color distribution as the picture content of the picture to be identified.
5. A modeling method, the method comprising:
acquiring a sample picture of known content information;
analyzing the sample picture to determine the color distribution thereof;
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
6. The method of claim 5, wherein said analyzing said sample picture to determine its color distribution comprises:
identifying the color of the sample picture at each identification position based on the coordinate system and the color value;
determining the color and the shape of a color block in the picture to be recognized according to the color value of the picture to be recognized at each recognition position, or,
and determining the color and the shape of color blocks in the picture to be recognized and the relative positions of partial or all color blocks according to the color values of the picture to be recognized at all recognition positions.
7. The method of claim 6, wherein the statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture comprises:
counting the color distribution of a plurality of sample pictures under the same content information;
the statistically determined color value range and shape set are used as the reference color distribution of the same content information, or,
and taking the color value range, the shape set and the relative position set which are determined by statistics as the reference color distribution of the same content information.
8. A machine learning apparatus, comprising:
the acquisition module is used for acquiring a sample picture of known content information;
the analysis module is used for analyzing the sample picture to determine the color distribution of the sample picture;
a processing module to:
statistically determining a reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture, or,
and updating the reference color distribution corresponding to the content information based on the content information of the sample picture and the color distribution of the sample picture.
9. A picture content identification method is characterized by comprising the following steps:
analyzing the picture to be identified to determine the color distribution of the picture;
querying a material library to determine a combination of materials corresponding to the color distribution of the picture to be identified;
determining content information of the picture to be identified based on the combination of the materials.
10. An apparatus for recognizing picture contents, the apparatus comprising:
the color analysis module is used for analyzing the picture to be identified to determine the color distribution of the picture;
the combination determining module is used for inquiring a material library to determine the combination of materials corresponding to the color distribution of the picture to be identified;
a content determination module for determining content information of the picture to be identified based on the combination of the materials.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710448132.4A CN107292934B (en) | 2017-06-14 | 2017-06-14 | Picture content identification method, modeling method and related device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710448132.4A CN107292934B (en) | 2017-06-14 | 2017-06-14 | Picture content identification method, modeling method and related device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107292934A true CN107292934A (en) | 2017-10-24 |
| CN107292934B CN107292934B (en) | 2020-10-13 |
Family
ID=60096367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710448132.4A Expired - Fee Related CN107292934B (en) | 2017-06-14 | 2017-06-14 | Picture content identification method, modeling method and related device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107292934B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110555438A (en) * | 2019-07-19 | 2019-12-10 | 广东智媒云图科技股份有限公司 | machine-assisted learning method of basic language, terminal equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102819746A (en) * | 2012-07-10 | 2012-12-12 | 电子科技大学 | Method for automatically marking category of remote sensing image based on author-genre theme model |
| CN103927399A (en) * | 2014-05-07 | 2014-07-16 | 苏州搜客信息技术有限公司 | Information push method and device |
| CN104023249A (en) * | 2014-06-12 | 2014-09-03 | 腾讯科技（深圳）有限公司 | Method and device of identifying television channel |
| US20150189158A1 (en) * | 2013-12-28 | 2015-07-02 | Connor Hailey | Systems, methods, and media for presenting a field of view from a camera of a device on a display of the device in the background of non-camera functionality |
-
2017
- 2017-06-14 CN CN201710448132.4A patent/CN107292934B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102819746A (en) * | 2012-07-10 | 2012-12-12 | 电子科技大学 | Method for automatically marking category of remote sensing image based on author-genre theme model |
| US20150189158A1 (en) * | 2013-12-28 | 2015-07-02 | Connor Hailey | Systems, methods, and media for presenting a field of view from a camera of a device on a display of the device in the background of non-camera functionality |
| CN103927399A (en) * | 2014-05-07 | 2014-07-16 | 苏州搜客信息技术有限公司 | Information push method and device |
| CN104023249A (en) * | 2014-06-12 | 2014-09-03 | 腾讯科技（深圳）有限公司 | Method and device of identifying television channel |
Non-Patent Citations (4)
| Title |
|---|
| 刘振: "雾霾天气下交通标志的检测与识别", 《万方数据知识服务平台》 * |
| 吕军 等: "基于模板匹配的多目标水稻灯诱害虫识别方法的研究", 《中国水稻科学》 * |
| 季书芳: "基于区域特征及拓扑结构的图案识别", 《中国优秀博硕士学位论文全文数据库（硕士） 信息科技辑》 * |
| 徐胜: "三维物体识别研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110555438A (en) * | 2019-07-19 | 2019-12-10 | 广东智媒云图科技股份有限公司 | machine-assisted learning method of basic language, terminal equipment and storage medium |
| CN110555438B (en) * | 2019-07-19 | 2023-08-18 | 广东智媒云图科技股份有限公司 | Machine-assisted learning method for basic language, terminal equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107292934B (en) | 2020-10-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103971134B (en) | Image classification, retrieval and bearing calibration, and related device | |
| GB2391127A (en) | System and method for bounding and classifying regions within a graphical image. | |
| US20190236803A1 (en) | Pupil localizing method and system | |
| CN109344869A (en) | A kind of disaggregated model optimization method, device and storage equipment, program product | |
| WO2020005773A1 (en) | Food quality image classification | |
| JP2016167304A (en) | Method for improving classification result of classifier | |
| US20180314984A1 (en) | Retraining a machine classifier based on audited issue data | |
| CN112352412B (en) | Network traffic processing method and device, storage medium and computer equipment | |
| Łuszczkiewicz-Piątek | Which color space should be chosen for robust color image retrieval based on mixture modeling | |
| WO2017027031A1 (en) | Assigning classifiers to classify security scan issues | |
| CN107292934B (en) | Picture content identification method, modeling method and related device | |
| CN104462420A (en) | Method and device for executing query tasks on database | |
| CN109948653A (en) | A kind of image similarity judgment method, electronic equipment and storage medium | |
| US11302070B1 (en) | Systems and methods for multi-tree deconstruction and processing of point clouds | |
| CN109857897B (en) | Trademark image retrieval method and device, computer equipment and storage medium | |
| WO2017095439A1 (en) | Incremental clustering of a data stream via an orthogonal transform based indexing | |
| US11188648B2 (en) | Training a security scan classifier to learn an issue preference of a human auditor | |
| CN106294530B (en) | The method and system of rule match | |
| KR100312331B1 (en) | System and method for searching image based on contents | |
| US10248789B2 (en) | File clustering using filters working over file attributes | |
| CN110110127A (en) | A kind of method and electronic equipment of the primary color inks identifying spot color mixed ink | |
| Huang et al. | Key color priority based image recoloring for dichromats | |
| CN114745452B (en) | Equipment management method and device and electronic equipment | |
| LU504734B1 (en) | Self-supervised text recognition method, device and storage medium | |
| Naegel et al. | Colour image filtering with component-graphs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: Building N3, building 12, No. 27, Jiancai Chengzhong Road, Haidian District, Beijing 100096 Applicant after: Beijing Xingxuan Technology Co.,Ltd. Address before: 100085 Beijing, Haidian District on the road to the information on the ground floor of the 1 to the 3 floor of the 2 floor, room 11, 202 Applicant before: Beijing Xiaodu Information Technology Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201013 |