CN109872357A - A kind of article arrangement face accounting calculation method, system and electronic equipment - Google Patents
A kind of article arrangement face accounting calculation method, system and electronic equipment Download PDFInfo
- Publication number
- CN109872357A CN109872357A CN201910041599.6A CN201910041599A CN109872357A CN 109872357 A CN109872357 A CN 109872357A CN 201910041599 A CN201910041599 A CN 201910041599A CN 109872357 A CN109872357 A CN 109872357A
- Authority
- CN
- China
- Prior art keywords
- pixel
- article
- arrangement face
- accounting
- article arrangement
- 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.)
- Pending
Links
- 238000004364 calculation method Methods 0.000 title claims abstract description 41
- 238000000034 method Methods 0.000 claims abstract description 44
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 238000000638 solvent extraction Methods 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims description 31
- 238000012545 processing Methods 0.000 claims description 17
- 230000011218 segmentation Effects 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 230000017105 transposition Effects 0.000 claims description 10
- 238000012512 characterization method Methods 0.000 claims description 7
- 238000010586 diagram Methods 0.000 description 10
- 238000005516 engineering process Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000013527 convolutional neural network Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 210000003127 knee Anatomy 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Landscapes
- Image Processing (AREA)
Abstract
The present invention provides a kind of article arrangement face accounting calculation method, comprising: step S1: obtaining images of items；Step S2: the corresponding pixel class of pixel in the images of items is obtained by image, semantic partitioning algorithm, pixel is divided into article pixel and background pixel by pixel class；Step S3: it is calculated according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting.The present invention also provides the systems and electronic equipment of the correspondence article arrangement face accounting calculation method, and the accounting in article arrangement face quick, intelligent can be measured by the method, system and electronic equipment.
Description
[technical field]
The present invention relates to deep learning field more particularly to a kind of article arrangement face accounting calculation methods, system and electronics
Equipment.
[background technique]
In article identification and detection field, we often need to carry out the calculating of article arrangement face accounting, as businessman needs to know
On road layer shelf when the accounting of current each article, then need to provide a kind of efficient, reliable article arrangement face accounting calculating side
Method calculates the arrangement plane accounting of each article.
Being applied at present in the deep learning algorithm of article identification and detection field is mainly object detection algorithms.Existing object
Product arrangement plane accounting calculation method usually passes through object detection algorithms and identifies single item generic, then calculates again each
The arrangement plane accounting of article.However, in order to train detection model used by object detection algorithms, it usually needs difference is individually
Data mark is carried out to object, this needs to expend a large amount of manpowers, and in the case where object intensively occurs and arbitrarily arranges, inspection
Model is surveyed to be difficult to carry out Accurate Prediction to appearance position.Such as in actual scene, article be most likely not by regularly placing, but
Intensively dispersedly occur.Under the scene of the bulk products such as refrigerator-freezer, stall identification, this point is especially apparent.In these applied fields
Jing Zhong, existing article arrangement face accounting calculation method disadvantage are more obvious.
[summary of the invention]
To overcome the problems of the prior art, the present invention provides a kind of article arrangement face accounting calculation method, system and electricity
Sub- equipment.
The present invention provides a kind of technical solutions for solving above-mentioned technical problem: a kind of article arrangement face accounting calculating side
Method, article arrangement face accounting calculation method include: step S1: obtaining images of items；Step S2: pass through image, semantic partitioning algorithm
The corresponding pixel class of pixel in the images of items is obtained, pixel is divided into article pixel and background pixel by pixel class；And
Step S3: it is calculated according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting.
Preferably, between step S2 and S3 include step Sa: to article pixel region carry out expansion process so that
Few etch away parts background pixel.
Preferably, described image semantic segmentation algorithm includes: step S21: carrying out down-sampled acquisition low-dimensional to images of items
The downscaled images of characterization；And step S22: downscaled images are up-sampled to obtain the enlarged drawing for restoring dimension, and are obtained
The corresponding pixel class of the pixel of enlarged drawing, the size of enlarged drawing are consistent with the images of items.
Preferably, the article pixel includes the N seed article pixel corresponding to N kind difference article, the article arrangement
Face accounting includes the arrangement plane accounting of N kind difference article, and N is more than or equal to 1.
Preferably, in step Sa, the article pixel carries out expansion process with etch away parts background pixel or whole
Background pixel.
Preferably, the expansion process includes: step S31: will be converted to bianry image by the image of step S2 processing；
Step S32: one operator of selection；And step S33: at least partly article pixel is traversed by operator and carries out expansive working acquisition newly
Bianry image.
The present invention also provides a kind of article arrangement face accounting computing systems, are used to calculate article arrangement face accounting, article
Arrangement plane accounting computing system includes: input module；For obtaining images of items；Image, semantic dividing processing module；For leading to
It crosses image, semantic partitioning algorithm and obtains the corresponding pixel class of pixel in the images of items, pixel is divided into article by pixel class
Pixel and background pixel；And accounting computing module；Article row is obtained for calculating according to the corresponding elemental area of each pixel class
Column face accounting.
Preferably, article arrangement face accounting computing system further comprises: expansion process module；For to image, semantic point
It cuts the article pixel obtained in processing module and carries out expansion process at least etch away parts background pixel；Accounting computing module root
It is calculated according to the corresponding elemental area of pixel Jing Guo the output of expansion process module and obtains article arrangement face accounting.
Preferably, described image semantic segmentation processing module is FCN model comprising convolution module and transposition convolution mould
Block；Convolution module is used to carry out images of items the downscaled images of down-sampled acquisition low-dimensional characterization；And transposition convolution module is used for
Downscaled images are up-sampled to obtain the enlarged drawing for restoring dimension, and the corresponding pixel class of pixel for obtaining enlarged drawing
Not, the size of enlarged drawing is consistent with the images of items.
The present invention provides a kind of electronic equipment, including memory and processor, is stored with computer journey in the memory
Sequence, the computer program are arranged to execute article arrangement face accounting calculation method as described above when operation；The processing
Device is arranged to execute article arrangement face accounting calculation method as described above by the computer program.
Compared with prior art, article arrangement face accounting calculation method provided by the present invention passes through semantic segmentation algorithm meter
The pixel class for calculating pixel calculates according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting, due to nothing
Article number need to be calculated, entire operand greatly reduces, so that calculating quickly, intelligent and efficient arrangement plane accounting is calculated as
It may.
By semantic segmentation algorithm combination expansion process, expansion process is corroded background pixel, can be made most
The accounting obtained eventually is reduced closer to human eye visual sense such as gaps between articles bring article arrangement face accounting calculating side
Under method accounting obtained and human eye direct feel observed by account between error.
Article arrangement face accounting calculation method provided by the present invention can be allowed to fit by down-sampled and up-sampling
For different sizes, different types of image improves the versatility of article arrangement face accounting calculation method.
Expansion process operand is small, can efficiently corrode background pixel, reduces integral item arrangement plane accounting and calculates
The complexity of method promotes computational efficiency, reduces the hardware requirement for executing the device of the method.
Article arrangement face accounting computing system provided by the present invention and electronic equipment also have the advantages that described.
[Detailed description of the invention]
Fig. 1 is a kind of article arrangement face accounting calculation method schematic flow diagram of first embodiment of the invention.
Fig. 2 is a kind of embodiment schematic flow diagram of step S2 in first embodiment.
Fig. 3 is a kind of module diagram of article arrangement face accounting computing system of second embodiment of the invention.
Fig. 4 is a kind of article arrangement face accounting calculation method schematic flow diagram of third embodiment of the invention.
Fig. 5 is a kind of embodiment schematic flow diagram of step Sa in 3rd embodiment.
Fig. 6 is a kind of module diagram of article arrangement face accounting computing system of fourth embodiment of the invention.
Fig. 7 is the module diagram of fifth embodiment of the invention electronic equipment.
[specific embodiment]
In order to make the purpose of the present invention, technical solution and advantage are more clearly understood, below in conjunction with attached drawing and embodiment,
The present invention will be described in further detail.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention,
It is not intended to limit the present invention.
Referring to Fig. 1, first embodiment of the invention provides a kind of article arrangement face accounting calculation method 1, it is a kind of base
In the article arrangement face accounting calculation method of image, semantic identification, it is suitable for article arrangement face accounting and calculates.It is provided by the present invention
Article arrangement face accounting calculation method 1 be suitable for plurality of application scenes, for example, be arranged with various article on the shelf of refrigerator-freezer,
The accounting for obtaining all kinds of articles on shelf can be calculated by the article arrangement face accounting calculation method 1.In another example can lead to
It crosses the article arrangement face accounting calculation method 1 and calculates the arrangement plane accounting for obtaining each article in bulk product stockpile on ground.
Article arrangement face accounting calculation method 1 includes:
Step S1: images of items is obtained；
Step S2: the corresponding pixel class of pixel in the images of items, pixel are obtained by image, semantic partitioning algorithm
It is divided into article pixel and background pixel by pixel class；And
Step S3: it is calculated according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting.
In step sl, images of items pass through photographic device obtain, it is preferable that photographic device be set to article just on
Side, imaging area and article placement region are just corresponding, and further preferably, imaging area is the 90%- of article placement region
110%, the so error rate to reduce arrangement plane accounting.It include that item image information and background image are believed in images of items
Breath.It is appreciated that as being placed with laughable and two kinds of articles of Sprite on shelf, Item Information includes corresponding laughable and Sprite figure
As information, gap between adjacent cola, gap between gap or adjacent laughable and Sprite between adjacent Sprite and not
Placing the laughable region with Sprite part is background image information.
In step s 2, the corresponding pixel class of pixel in images of items, pixel are obtained using image, semantic partitioning algorithm
It is customized that classification can be user.Specific image, semantic partitioning algorithm includes but is not limited to Normalized cut (N-cut)
Technology, Grab cut technology, CNN (Convolutional Neural Network, convolutional neural networks) technology or FCN
(Fully convolutional networks, full convolutional neural networks) technology is realized, can also be implanted into the technology
Dilated Convolutions technology, condition random field (Conditional Random Field, CRF) technology and Ma Erke
Husband's random field (Markov Random Filed, MRF) technology etc..
Referring to Fig. 2, as one embodiment, the image, semantic partitioning algorithm in step S2 includes: step S21: to object
Product image carries out the downscaled images of down-sampled acquisition low-dimensional characterization；And
Step S22: to downscaled images up-sampled with obtain restore dimension enlarged drawing, and obtain enlarged drawing it
The corresponding pixel class of pixel, the size of enlarged drawing are consistent with the images of items.
As one embodiment, down-sampled in step S21 can be realized by convolution algorithm, be extracted by convolution algorithm
The different characteristic of images of items, convolution algorithm can realize that first layer convolutional layer can only be extracted by multilayer convolutional layer
The rudimentary feature such as levels such as edge, lines and angle, more convolutional layers can from low-level features the more complicated spy of iterative extraction
Sign.It is alternative in convolution algorithm to increase line rectification (Rectified Linear Units, ReLU) Ji Chihua (Pooling)
Processing.Up-sampling in step S22 is realized by transposition convolution, to obtain the enlarged drawing for restoring dimension, the ruler of enlarged drawing
Very little size is consistent with the images of items, and pixel and the images of items pixel of enlarged drawing map one by one, the pixel of enlarged drawing
Classification is also mapped in images of items pixel one by one.
As a kind of preferred embodiment, is calculated using FCN and obtain the corresponding pixel class of pixel in the images of items.
FCN can receive the images of items of arbitrary dimension as input, then by transposition convolutional layer to the spy of the last one convolutional layer
Sign figure (feature map) is up-sampled, and so that it is restored to the identical size of images of items, so as to each pixel
Produce the pixel class of a prediction.Meanwhile remain the spatial information in original goods image, finally with images of items
Etc. classify on the characteristic pattern of sizes to each pixel.
More than having the characteristics that due to FCN, the images of items size of FCN no matter is inputted, can be obtained by FCN
The corresponding pixel class information of each pixel into item pictures.
It is appreciated that pixel is divided into article pixel and background pixel by pixel class, article pixel includes corresponding to N kind not
With the N seed article pixel of article, N is more than or equal to 1.For being placed with laughable and two kinds of articles of Sprite on shelf above-mentioned,
It is article pixel by the corresponding pixel of pixel class Item Information, article pixel is different sons by the different instructions of specific article
Article pixel, it is the first sub- article pixel that such as cola is corresponding, and corresponding Sprite is the second sub- article pixel.It is carried on the back by pixel class
The corresponding pixel of scape image information is background pixel.
In step s3, article arrangement face accounting includes the arrangement plane accounting of N kind difference article, and N is more than or equal to 1.
For being placed with laughable and two kinds of articles of Sprite on shelf above-mentioned, it is assumed that obtained by semantic segmentation algorithm
Image be 64*64 size, wherein the first sub- article pixel be 1024, the second sub- article pixel be 3008, background pixel
It is 64；I.e. first sub- article pixel accounts for the 25% of image, and the second sub- article pixel accounts for about the 73.5% of image, and background pixel is about
Account for 1.5%.We can think that arrangement plane accounting laughable on shelf is 25%, and the arrangement plane accounting of Sprite is 73.5%,
The arrangement plane accounting of background is 1.5%.
Referring to Fig. 3, second embodiment of the invention, which provides one kind, corresponds to the calculating of first embodiment article arrangement face accounting
The article arrangement face accounting computing system 20 of method, is used to calculate article arrangement face accounting.Article arrangement face accounting calculates system
System 20 includes:
Input module 21；For obtaining images of items；
Image, semantic dividing processing module 22；For obtaining pixel in the images of items by image, semantic partitioning algorithm
Corresponding pixel class, pixel are divided into article pixel and background pixel by pixel class；And
Accounting computing module 24；It is accounted for for calculating acquisition article arrangement face according to the corresponding elemental area of each pixel class
Than.
Preferably, described image semantic segmentation processing module is FCN model comprising convolution module and transposition convolution module
(not shown)；Convolution module is used to carry out images of items the downscaled images of down-sampled acquisition low-dimensional characterization；Transposition convolution mould
Block is used to up-sample to obtain the enlarged drawing for restoring dimension downscaled images, and the pixel for obtaining enlarged drawing is corresponding
Pixel class, the size of enlarged drawing are consistent with the images of items.
It is understood that content disclosed in first embodiment is suitable for the present embodiment.
Third embodiment of the invention provides a kind of article arrangement face accounting calculation method 10, is provided with first embodiment
A kind of article arrangement face accounting calculation method 1 the difference is that only increase step Sa, make in step S3 calculate obtain
Visual sense of the arrangement plane accounting obtained closer to human eye.Specifically, article arrangement face accounting calculation method 10 includes:
Step S1: images of items is obtained；
Step S2: the corresponding pixel class of pixel in the images of items, pixel are obtained by image, semantic partitioning algorithm
It is divided into article pixel and background pixel by pixel class；
Step Sa: expansion process is carried out at least etch away parts background pixel to article pixel region；And
Step S3: it is calculated according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting.
Step S1 and step S2 is identical as step S1 in first embodiment and step S2, and details are not described herein again.
In step Sa, the article pixel carries out expansion process with etch away parts background pixel or whole background pictures
Element.As one embodiment, step Sa uses bianry image expansive working, and step Sa can be specifically included:
Step Sa1: bianry image will be converted to by the image of step S2 processing；
Step Sa2: one operator of selection；And
Step Sa3: the new bianry image for carrying out expansive working acquisition is traversed at least partly article pixel by operator.
It is appreciated that the operator is structural element.Step Sa3 iteration can be carried out according to actual needs to obtain
Preferable expansion process effect.Preferably, in refrigerator-freezer application scenarios, step Sa3 is repeated 4-8 times, that is, it is swollen to repeat iteration
Swollen operation 4-8 times.
In step s3, article arrangement face accounting includes the arrangement plane accounting of N kind difference article, and N is more than or equal to 1.Work as step
In rapid Sa, the article pixel carry out expansion process only etch away parts background pixel when, article arrangement face accounting is also wrapped
Include the arrangement plane accounting of background.
For being placed with laughable and two kinds of articles of Sprite on shelf above-mentioned, it is assumed that pass through expansion process obtained two
Value image is 64*64 size, and background pixel is all corroded, and the first sub- article pixel is 1024 in bianry image, the second son
Article pixel is 3072, i.e., the accounting of the first sub- article pixel and the second sub- article pixel is 1:3, then the first sub- article picture
Element accounts for the 25% of bianry image, and the second sub- article pixel accounts for the 75% of bianry image, it can thinks arrangement laughable on shelf
Face accounting is 25%, and the arrangement plane accounting of Sprite is 75%.It is compared with same example in first embodiment, first embodiment
The background pixel accounting calculated in example is 1.5%, which is likely to be present between Sprite and cola and unavoidably deposits
Gap brought by, but actually shelf are filled up by article, for the direct feel of human eye, the background pixel
Accounting should be ignored.And this problem can be well solved by expansion process, so that calculating the accounting obtained more
Close to the direct feel of human eye.
In another example, since shelf have a large amount of positions not filled up by article, background pixel only portion in expansion process
Divide and be corroded, it is assumed that passing through expansion process bianry image obtained is 64*64 size, the first sub- article pixel in bianry image
It is 1024, the second sub- article pixel is 2048, and background pixel is 1024.I.e. first sub- article pixel and the second sub- article
Pixel and the accounting of background pixel are 1:2:1, then the first sub- article pixel accounts for the 25% of bianry image, the second sub- article picture
Element accounts for the 50% of bianry image, and background pixel accounts for 25%.It can think that arrangement plane accounting laughable on shelf is 25%, snow
Green arrangement plane accounting is 50%, and the arrangement plane accounting of background is 25%.
Fourth embodiment of the invention provides a kind of article corresponding to 3rd embodiment article arrangement face accounting calculation method
Arrangement plane accounting computing system 40 is used to calculate article arrangement face accounting.Article arrangement face accounting computing system 40 includes:
Input module 41；For obtaining images of items；
Image, semantic dividing processing module 42；For obtaining pixel in the images of items by image, semantic partitioning algorithm
Corresponding pixel class, pixel are divided into article pixel and background pixel by pixel class；
Expansion process module 43；For being carried out at expansion to the article pixel obtained in image, semantic dividing processing module 43
Reason is at least etch away parts background pixel；And
Accounting computing module 44；It is accounted for for calculating acquisition article arrangement face according to the corresponding elemental area of each pixel class
Than.
Preferably, described image semantic segmentation processing module is FCN model comprising convolution module and transposition convolution module
(not shown)；Convolution module is used to carry out images of items the downscaled images of down-sampled acquisition low-dimensional characterization；Transposition convolution mould
Block is used to up-sample to obtain the enlarged drawing for restoring dimension downscaled images, and the pixel for obtaining enlarged drawing is corresponding
Pixel class, the size of enlarged drawing are consistent with the images of items.
Preferably, operator is stored in the expansion process module；The expansion process module includes bianry image conversion
Module and expansive working module (not shown)；Bianry image conversion module will be for that will pass through at image, semantic dividing processing module
The image of reason is converted to bianry image；Expansive working module is used to expand at least partly article pixel traversal by operator
Operation obtains new bianry image.
It is understood that content disclosed in 3rd embodiment is suitable for the present embodiment.
Fifth embodiment of the invention provides a kind of electronic equipment 50, including memory 52 and processor 51, the memory
Computer program is stored in 52, the computer program is arranged to execute the row of article described in first embodiment when operation
Column face accounting calculation method.The processor 51 is arranged to execute object described in first embodiment by the computer program
Product arrangement plane accounting calculation method.
Electronic equipment 50 can be with display screen and support the various electronic equipments of video playing, including but not limited to
Smart phone, tablet computer, E-book reader, MP3 player (Moving Picture Experts Group Audio
Layer III, dynamic image expert's compression standard audio level 3), MP4 (Moving Picture Experts Group
Audio Layer IV, dynamic image expert's compression standard audio level 4) player, pocket computer on knee and desk-top meter
Calculation machine etc..
Compared with prior art, article arrangement face accounting calculation method provided by the present invention passes through semantic segmentation algorithm meter
The pixel class for calculating pixel calculates according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting, due to nothing
Article number need to be calculated, entire operand greatly reduces, so that calculating quickly, intelligent and efficient arrangement plane accounting is calculated as
It may.
By semantic segmentation algorithm combination expansion process, expansion process is corroded background pixel, can be made most
The accounting obtained eventually is reduced closer to human eye visual sense such as gaps between articles bring article arrangement face accounting calculating side
Under method accounting obtained and human eye direct feel observed by account between error.
Article arrangement face accounting calculation method provided by the present invention can be allowed to fit by down-sampled and up-sampling
For different sizes, different types of image improves the versatility of article arrangement face accounting calculation method.
Expansion process operand is small, can efficiently corrode background pixel, reduces integral item arrangement plane accounting and calculates
The complexity of method promotes computational efficiency, reduces the hardware requirement for executing the device of the method.
Article arrangement face accounting computing system provided by the present invention and electronic equipment also have the advantages that described.
Disclosed embodiment according to the present invention may be implemented as computer software above with reference to the process of flow chart description
Program.For example, embodiment disclosed by the invention includes a kind of computer program product comprising be carried on computer storage medium
On computer program, which includes the program code for method shown in execution flow chart.In such reality
It applies in example, which can be downloaded and installed from network by communications portion, and/or be pacified from detachable media
Dress.When the computer program is executed by processor, the above-mentioned function of limiting in the present processes is executed.It needs to illustrate
Be, computer storage described herein can be computer-readable signal media or computer readable storage medium or
It is the two any combination.Computer readable storage medium for example may be-but not limited to-electricity, magnetic, light, electricity
Magnetic, the system of infrared ray or semiconductor, device or device, or any above combination.Computer readable storage medium is more
Specific example can include but is not limited to: have the electrical connections of one or more conducting wires, portable computer diskette, hard disk,
Random access storage device (RAM), read-only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber,
Portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory device or above-mentioned any appropriate group
It closes.In this application, it includes or the tangible medium of storage program that the program can be with that computer readable storage medium, which can be any,
It is commanded execution system, device or device use or in connection.And in this application, computer-readable signal
Medium may include in a base band or as the data-signal that carrier wave a part is propagated, wherein carrying computer-readable journey
Sequence code.The data-signal of this propagation can take various forms, including but not limited to electromagnetic signal, optical signal or above-mentioned
Any appropriate combination.Computer-readable signal media can also be any computer other than computer readable storage medium
Readable medium, the computer-readable medium can be sent, propagated or transmitted for by instruction execution system, device or device
Using or program in connection.The program code for including on computer-readable medium can be with any suitable medium
Transmission, including but not limited to: wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The calculating of the operation for executing the application can be write with one or more programming languages or combinations thereof
Machine program code, described program design language include object oriented program language-such as Java, Smalltalk, C+
+, it further include conventional procedural programming language-such as " C " language or similar programming language.Program code can
Fully to execute, partly execute on the user computer on the user computer, be executed as an independent software package,
Part executes on the remote computer or executes on a remote computer or server completely on the user computer for part.
In situations involving remote computers, remote computer can pass through the network of any kind --- including local area network (LAN)
Or wide area network (WAN)-is connected to subscriber computer, or, it may be connected to outer computer (such as utilize Internet service
Provider is connected by internet).
Flow chart and module map in attached drawing illustrate system, method and computer according to the various embodiments of the application
The architecture, function and operation in the cards of program product.In this regard, each box in flowchart or block diagram can be with
A part of a module, program segment or code is represented, a part of the module, program segment or code includes one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, or can be referred to specialized hardware and computer
The combination of order is realized.
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all in original of the invention
Made any modification within then, equivalent replacement and improvement etc. should all be comprising within protection scope of the present invention.
Claims (10)
1. a kind of article arrangement face accounting calculation method, it is characterised in that: article arrangement face accounting calculation method includes:
Step S1: images of items is obtained；
Step S2: the corresponding pixel class of pixel in the images of items is obtained by image, semantic partitioning algorithm, pixel presses picture
Plain classification is divided into article pixel and background pixel；And
Step S3: it is calculated according to the corresponding elemental area of each pixel class and obtains article arrangement face accounting.
2. article arrangement face accounting calculation method as described in claim 1, it is characterised in that: include between step S2 and S3
Step Sa: expansion process is carried out at least etch away parts background pixel to article pixel region.
3. article arrangement face accounting calculation method as described in claim 1, it is characterised in that: described image semantic segmentation algorithm
Include:
Step S21: the downscaled images of down-sampled acquisition low-dimensional characterization are carried out to images of items；And
Step S22: downscaled images are up-sampled to obtain the enlarged drawing for restoring dimension, and obtain the pixel of enlarged drawing
Corresponding pixel class, the size of enlarged drawing are consistent with the images of items.
4. article arrangement face accounting calculation method as described in claim 1, it is characterised in that: the article pixel includes corresponding to
In the N seed article pixel of N kind difference article, article arrangement face accounting includes the arrangement plane accounting of N kind difference article, N
More than or equal to 1.
5. article arrangement face accounting calculation method as claimed in claim 2, it is characterised in that: in step Sa, the article
Pixel carries out expansion process with etch away parts background pixel or whole background pixels.
6. article arrangement face accounting calculation method as described in claim 1, it is characterised in that: the expansion process includes:
Step S31: bianry image will be converted to by the image of step S2 processing；
Step S32: one operator of selection；And
Step S33: progress expansive working is traversed at least partly article pixel by operator and obtains new bianry image.
7. a kind of article arrangement face accounting computing system, is used to calculate article arrangement face accounting, it is characterised in that: article arrangement
Face accounting computing system includes:
Input module；For obtaining images of items；
Image, semantic dividing processing module；It is corresponding for obtaining pixel in the images of items by image, semantic partitioning algorithm
Pixel class, pixel are divided into article pixel and background pixel by pixel class；And
Accounting computing module；Article arrangement face accounting is obtained for calculating according to the corresponding elemental area of each pixel class.
8. article arrangement face accounting computing system as claimed in claim 7, it is characterised in that: article arrangement face accounting calculates system
System further comprises:
Expansion process module；For carrying out expansion process to the article pixel obtained in image, semantic dividing processing module at least
Etch away parts background pixel；Accounting computing module is according to the corresponding each pixel class of pixel Jing Guo the output of expansion process module
Elemental area calculate obtain article arrangement face accounting.
9. article arrangement face accounting computing system as claimed in claim 7, it is characterised in that: the processing of described image semantic segmentation
Module is FCN model comprising convolution module and transposition convolution module；
Convolution module is used to carry out images of items the downscaled images of down-sampled acquisition low-dimensional characterization；And
Transposition convolution module is used to up-sample to obtain the enlarged drawing for restoring dimension downscaled images, and obtains enlarged drawing
As the corresponding pixel class of pixel, the size of enlarged drawing is consistent with the images of items.
10. a kind of electronic equipment, including memory and processor, it is characterised in that: be stored with computer journey in the memory
Sequence, the computer program are arranged to perform claim when operation and require article arrangement face accounting meter described in any one of 1-6
Calculation method；
The processor is arranged to require article arrangement described in any one of 1-6 by the computer program perform claim
Face accounting calculation method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910041599.6A CN109872357A (en) | 2019-01-16 | 2019-01-16 | A kind of article arrangement face accounting calculation method, system and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910041599.6A CN109872357A (en) | 2019-01-16 | 2019-01-16 | A kind of article arrangement face accounting calculation method, system and electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109872357A true CN109872357A (en) | 2019-06-11 |
Family
ID=66917787
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910041599.6A Pending CN109872357A (en) | 2019-01-16 | 2019-01-16 | A kind of article arrangement face accounting calculation method, system and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109872357A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782464A (en) * | 2019-11-04 | 2020-02-11 | 浙江大华技术股份有限公司 | Calculation method of object accumulation 3D space occupancy rate, coder-decoder and storage device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709568A (en) * | 2016-12-16 | 2017-05-24 | 北京工业大学 | RGB-D image object detection and semantic segmentation method based on deep convolution network |
| CN108427951A (en) * | 2018-02-08 | 2018-08-21 | 腾讯科技（深圳）有限公司 | Image processing method, device, storage medium and computer equipment |
| CN108829826A (en) * | 2018-06-14 | 2018-11-16 | 清华大学深圳研究生院 | A kind of image search method based on deep learning and semantic segmentation |
| CN108846795A (en) * | 2018-05-30 | 2018-11-20 | 北京小米移动软件有限公司 | Image processing method and device |
| CN108921163A (en) * | 2018-06-08 | 2018-11-30 | 南京大学 | A kind of packaging coding detection method based on deep learning |
| CN109086770A (en) * | 2018-07-25 | 2018-12-25 | 成都快眼科技有限公司 | A kind of image, semantic dividing method and model based on accurate scale prediction |
-
2019
- 2019-01-16 CN CN201910041599.6A patent/CN109872357A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106709568A (en) * | 2016-12-16 | 2017-05-24 | 北京工业大学 | RGB-D image object detection and semantic segmentation method based on deep convolution network |
| CN108427951A (en) * | 2018-02-08 | 2018-08-21 | 腾讯科技（深圳）有限公司 | Image processing method, device, storage medium and computer equipment |
| CN108846795A (en) * | 2018-05-30 | 2018-11-20 | 北京小米移动软件有限公司 | Image processing method and device |
| CN108921163A (en) * | 2018-06-08 | 2018-11-30 | 南京大学 | A kind of packaging coding detection method based on deep learning |
| CN108829826A (en) * | 2018-06-14 | 2018-11-16 | 清华大学深圳研究生院 | A kind of image search method based on deep learning and semantic segmentation |
| CN109086770A (en) * | 2018-07-25 | 2018-12-25 | 成都快眼科技有限公司 | A kind of image, semantic dividing method and model based on accurate scale prediction |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110782464A (en) * | 2019-11-04 | 2020-02-11 | 浙江大华技术股份有限公司 | Calculation method of object accumulation 3D space occupancy rate, coder-decoder and storage device |
| CN110782464B (en) * | 2019-11-04 | 2022-07-15 | 浙江大华技术股份有限公司 | Calculation method of object accumulation 3D space occupancy rate, coder-decoder and storage device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108090916B (en) | Method and apparatus for tracking the targeted graphical in video | |
| US10810435B2 (en) | Segmenting objects in video sequences | |
| CN108229341B (en) | Classification method and device, electronic equipment and computer storage medium | |
| CN108846440B (en) | Image processing method and device, computer readable medium and electronic equipment | |
| CN108510472B (en) | Method and apparatus for handling image | |
| CN108307240B (en) | Video recommendation method and device | |
| CN108062780A (en) | Method for compressing image and device | |
| CN109117831A (en) | The training method and device of object detection network | |
| CN107911753A (en) | Method and apparatus for adding digital watermarking in video | |
| CN108197623A (en) | For detecting the method and apparatus of target | |
| CN110163215A (en) | Image processing method, device, computer-readable medium and electronic equipment | |
| CN108491816A (en) | The method and apparatus for carrying out target following in video | |
| CN108446651A (en) | Face identification method and device | |
| CN109934181A (en) | Text recognition method, device, equipment and computer-readable medium | |
| CN109165573A (en) | Method and apparatus for extracting video feature vector | |
| CN108230346B (en) | Method and device for segmenting semantic features of image and electronic equipment | |
| CN108986169A (en) | Method and apparatus for handling image | |
| CN110414502A (en) | Image processing method and device, electronic equipment and computer-readable medium | |
| CN109690471A (en) | Use the media hype of orientation metadata | |
| EP4432215A1 (en) | Image processing method and device | |
| CN110309720A (en) | Video detecting method, device, electronic equipment and computer-readable medium | |
| CN108090885A (en) | For handling the method and apparatus of image | |
| CN110427915B (en) | Method and apparatus for outputting information | |
| CN109389096A (en) | Detection method and device | |
| CN108898604A (en) | Method and apparatus for handling image |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190611 |