CN104851183A - Paper currency face and orientation recognition method and device - Google Patents
Paper currency face and orientation recognition method and device Download PDFInfo
- Publication number
- CN104851183A CN104851183A CN201510250574.9A CN201510250574A CN104851183A CN 104851183 A CN104851183 A CN 104851183A CN 201510250574 A CN201510250574 A CN 201510250574A CN 104851183 A CN104851183 A CN 104851183A
- Authority
- CN
- China
- Prior art keywords
- numbers
- ordered series
- gray scale
- bank note
- region
- 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
Abstract
The invention discloses a paper currency face and orientation recognition method and a device. The method comprises steps: a paper currency image of to-be-recognized paper currency is acquired; according to a set division rule, the paper currency image is divided into a first set number of regions; the gray sum of all regions in the paper currency image is calculated, and a gray sum sequence with a first set number is formed; the sequence is inputted to a recognizer in a BP neural network for recognition so as to obtain the face and the orientation of the paper currency, wherein network parameters of the recognizer of the BP neural network are obtained through training. Recognition of the face and the orientation of the paper currency of different currency value versions can be realized, only one training needs to be carried out on the face and the orientation of the paper currency of different currency value versions, the network parameters can be obtained, the face and the orientation of each currency value version can be recognized universally, and as the BP neural network recognition mode is adopted, the recognition efficiency of the face and the orientation of the paper currency is improved.
Description
Technical field
The embodiment of the present invention relates to paper money recognition technology, particularly relates to a kind of bank note towards recognition methods and device.
Background technology
In the automatic identifying of bank note, bank note towards be all follow-up identify judge bases, if bank note version is towards identification error, will directly cause the flase drop of all recognizers below, therefore bank note is basis and necessary identification step towards being identified in paper money recognition process.
In prior art, bank note is carry out extracting and identifying according to the feature of bank note towards identification, value of money version for different bank note needs to carry out different feature extractions, identification division also needs to revise accordingly according to the difference of value of money version etc., when collection environment change, (as brightness change etc.) also can cause the amendment carrying out corresponding technical scheme simultaneously, therefore causes bank note towards the not versatility identified.
Summary of the invention
In view of this, the embodiment of the present invention provides a kind of bank note towards recognition methods and device, with each value of money version of general identification towards.
First aspect, embodiments provide a kind of bank note towards recognition methods, described method comprises:
Obtain the banknote image of bank note to be identified;
According to setting division rule, described banknote image is divided into the region of the first setting quantity;
Calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity;
The recognizer of described ordered series of numbers input BP neural network is identified, with identify obtain bank note to be identified towards, wherein, the network parameter of the recognizer of described BP neural network is obtained by training.
Further, before by the recognizer of described ordered series of numbers input BP neural network, also comprise:
Described ordered series of numbers is normalized, obtains normalization ordered series of numbers.
Further, according to setting division rule, the region described banknote image being divided into the first setting quantity comprises:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
Further, calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity, comprising:
In units of pixel in each region, carry out the pixel gradation intervals sampling setting step-length;
Calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
Further, also comprise:
To multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version of bank note to be identified;
Adopt BP neural network to described sample training, obtain the network parameter of the recognizer of BP neural network.
Second aspect, the embodiment of the present invention additionally provides a kind of bank note towards recognition device, and described device comprises:
Banknote image acquisition module, for obtaining the banknote image of bank note to be identified;
Image-region divides module, for according to setting division rule, described banknote image is divided into the region of the first setting quantity;
Gray scale and ordered series of numbers form module, for calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity;
Bank note towards identification module, for the recognizer of described ordered series of numbers input BP neural network is identified, with identify obtain bank note to be identified towards, wherein, the network parameter of the recognizer of described BP neural network is obtained by training.
Further, also comprise:
Ordered series of numbers normalization module, for before the recognizer described ordered series of numbers being inputted BP neural network, is normalized described ordered series of numbers, obtains normalization ordered series of numbers.
Further, described image-region divide module specifically for:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
Further, described gray scale and ordered series of numbers formation module comprise:
Pixel sampling unit, in units of the pixel in each region, carries out the pixel gradation intervals sampling setting step-length;
Gray scale and computing unit, calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
Further, also comprise:
Module chosen by gray scale and ordered series of numbers sample, for multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version to bank note to be identified;
Network parameter training module, for adopting BP neural network to described sample training, obtains the network parameter of the recognizer of BP neural network.
The bank note that the embodiment of the present invention provides is towards recognition methods and device, by obtaining the banknote image of bank note to be identified, according to setting division rule, described banknote image is divided into the region of the number with the first setting quantity, and calculate each region gray scale and, formed a gray scale with the first setting quantity and ordered series of numbers, described ordered series of numbers input is had in the recognizer of the BP neural network of the network parameter of having trained and identifies, obtain bank note to be identified towards, achieve to the bank note of different value of money version towards identification, as long as obtain network parameter to the bank note of different value of money version towards respectively once training, just each value of money version of identification that can be general towards, the mode adopting BP neural network to carry out identifying improve bank note towards recognition efficiency.
Accompanying drawing explanation
Fig. 1 is the process flow diagram of a kind of bank note of providing of the embodiment of the present invention one towards recognition methods;
Fig. 2 is the process flow diagram of a kind of bank note of providing of the embodiment of the present invention two towards recognition methods;
Fig. 3 is the schematic diagram of a kind of bank note of providing of the embodiment of the present invention three towards recognition device.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described in further detail.Be understandable that, specific embodiment described herein is only for explaining the present invention, but not limitation of the invention.It also should be noted that, for convenience of description, illustrate only part related to the present invention in accompanying drawing but not full content.
Embodiment one
Fig. 1 is the process flow diagram of a kind of bank note of providing of the embodiment of the present invention one towards recognition methods, the present embodiment is applicable to ATM (Automatic Teller Machine, ATM (Automatic Teller Machine)) on to bank note towards identification, the method can be performed by ATM, specifically comprises the steps:
Step 110, obtains the banknote image of bank note to be identified.
The banknote image of bank note to be identified is gathered by image collecting device (as first-class in made a video recording).
Step 120, according to setting division rule, is divided into the region of the first setting quantity by described banknote image.
Wherein, described setting division rule defines and how to divide banknote image, namely divides the mode of banknote image, and the quantity defining the region formed after banknote image divides is the first setting quantity.Wherein, described first setting quantity can be 20.
Step 130, calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
The gray scale of pixel in described banknote image is obtained by the banknote image of bank note to be identified, by calculating the gray scale sum of the pixel in each region in described banknote image, obtain each region in described banknote image gray scale and, by the gray scale in region each in described banknote image be stored in an ordered series of numbers according to zone sequence, form gray scale and ordered series of numbers that one has the first setting quantity.
Step 140, identifies the recognizer of described ordered series of numbers input BP neural network, with identify obtain bank note to be identified towards.
Wherein, BP (Back Propagation) neural network is proposed by the scientist group headed by Rumelhart and McCelland for 1986, being a kind of Multi-layered Feedforward Networks by Back Propagation Algorithm training, is one of current most widely used neural network model.BP network can learn and store a large amount of input-output mode map relations, and without the need to disclosing the math equation describing this mapping relations in advance.Its learning rules use method of steepest descent, constantly adjusted the weights and threshold of network, make the error sum of squares of network minimum by backpropagation.BP neural network model topological structure comprises input layer, hidden layer and output layer.
The weights and threshold of described network parameter and BP neural network, comprising: input layer is to the weights and threshold of hidden layer, the weights and threshold of hidden layer to output layer and the number of plies of hidden layer.
Gray scale step 130 formed and ordered series of numbers are input in the recognizer of BP neural network and identify, just can obtain bank note to be identified towards.Wherein, the network parameter of the recognizer of described BP neural network is obtained by training.When described BP neural network is trained, to each bank note towards sample carry out feature extraction according to step 120-step 130.
The present embodiment is by obtaining the banknote image of bank note to be identified, according to setting division rule, described banknote image is divided into the region of the number with the first setting quantity, and calculate each region gray scale and, formed a gray scale with the first setting quantity and ordered series of numbers, described ordered series of numbers input is had in the recognizer of the BP neural network of the network parameter of having trained and identifies, obtain bank note to be identified towards, achieve to the bank note of different value of money version towards identification, as long as obtain network parameter to the bank note of different value of money version towards respectively once training, just each value of money version of identification that can be general towards, the mode adopting BP neural network to carry out identifying improve bank note towards recognition efficiency.
On the basis of technique scheme, according to setting division rule, described banknote image is divided into the region of the first setting quantity, preferably includes:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
According to the pixel matrix form of described banknote image, described banknote image is carried out decile, be divided into have the first setting quantity etc. subregion, wherein, described first setting quantity can be 20, namely the wide of described banknote image can be divided into 5, the height of described banknote image is divided into 4, thus described banknote image is divided into the subregions such as 20.Decile is carried out to banknote image, is beneficial to follow-up process, improve treatment effeciency.
On the basis of technique scheme, calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity, preferably include:
In units of pixel in each region, carry out the pixel gradation intervals sampling setting step-length;
Calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
Wherein, described setting procedure is 5 pixels.The sampling of pixel gradation intervals is carried out to each region, the resolution in each region can be reduced, reduce data volume, be conducive to improving data-handling efficiency.
On the basis of technique scheme, training and operation can be completed in advance, namely also preferably include:
To multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version of bank note to be identified;
Adopt BP neural network to described sample training, obtain the network parameter of the recognizer of BP neural network.
Any one bank note towards there being four kinds of situations, namely front forward, front oppositely, reverse side forward and reverse side reverse.To each gray scale and ordered series of numbers sample towards choosing the second setting quantity respectively of the value of money version of bank note to be identified, adopting BP neural network to described sample training, obtaining the network parameter of the recognizer of BP neural network.Wherein, described second setting quantity be greater than or equal to 100 integer.Each sample training towards the second setting quantity can be chosen, each sample carries out process according to the mode of the step 110-step 130 in embodiment one and obtains gray scale and ordered series of numbers sample, and formation input feature vector value matrix combined by all gray scales and sample.Utilize BP neural network to train this input feature vector value matrix, obtain the network parameter of the recognizer of the BP neural network of having trained.
Embodiment two
Fig. 2 be a kind of bank note of providing of the embodiment of the present invention two towards the process flow diagram of recognition methods, specifically comprise the steps:
Step 210, obtains the banknote image of bank note to be identified.
Step 220, according to setting division rule, is divided into the region of the first setting quantity by described banknote image.
Step 230, calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
Step 240, is normalized described ordered series of numbers, obtains normalization ordered series of numbers.
From step 230, the data representation gray scale in described ordered series of numbers.First the maximum gradation value in described ordered series of numbers is searched, by other gray-scale values except maximum gradation value in described ordered series of numbers divided by described maximum gradation value, change described maximum gradation value into 1 again, keep the invariant position of each gray-scale value in ordered series of numbers, thus obtaining normalization ordered series of numbers, the data in described normalization ordered series of numbers are the relative gray values after normalization.By being normalized gray scale and ordered series of numbers, can automatically adapt to bank note scene, avoiding susceptibility when bank note brightness and defect etc. are identified.
Step 250, identifies the recognizer of described ordered series of numbers input BP neural network, with identify obtain bank note to be identified towards.
Identify in the recognizer of the ordered series of numbers input BP neural network after the normalization that step 240 is obtained, with identify obtain bank note to be identified towards.Wherein, the network parameter of the recognizer of described BP neural network is obtained by training.When described BP neural network is trained, to each bank note towards sample carry out feature extraction according to step 210-step 240.
The present embodiment is by being divided into the region of the number with the first setting quantity by the banknote image of bank note to be identified according to setting division rule, calculate each region gray scale and, by the gray scale in all regions of described banknote image and combining, form gray scale and ordered series of numbers that one has the first setting quantity, and described ordered series of numbers is normalized, obtain the ordered series of numbers after normalization, identify in the recognizer of the ordered series of numbers input BP neural network after normalization, just can identify obtain bank note to be identified towards, achieve to the bank note of different value of money version towards identification, as long as obtain network parameter to the bank note of different value of money version towards respectively once training, just each value of money version of identification that can be general towards, the mode adopting BP neural network to carry out identifying improve bank note towards recognition efficiency, compared with embodiment one, the present embodiment has carried out normalization in characteristic extraction part to gray scale and ordered series of numbers, various bank note scene can be adapted to, avoid susceptibility when bank note brightness and defect etc. are identified.
Embodiment three
Fig. 3 is the schematic diagram of a kind of bank note of providing of the embodiment of the present invention three towards recognition device, as shown in Figure 3, the bank note that the present embodiment provides comprises towards recognition device: banknote image acquisition module 310, image-region division module 320, gray scale and ordered series of numbers formation module 330 and bank note are towards identification module 340.
Wherein, banknote image acquisition module 310 is for obtaining the banknote image of bank note to be identified;
Image-region divides module 320 for according to setting division rule, described banknote image is divided into the region of the first setting quantity;
Gray scale and ordered series of numbers formed module 330 for calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity;
Bank note towards identification module 340 for the recognizer of described ordered series of numbers input BP neural network is identified, with identify obtain bank note to be identified towards, wherein, the network parameter of the recognizer of described BP neural network is obtained by training.
Preferably, also comprise:
Ordered series of numbers normalization module, for before the recognizer described ordered series of numbers being inputted BP neural network, is normalized described ordered series of numbers, obtains normalization ordered series of numbers.
Preferably, described image-region divide module specifically for:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
Preferably, described gray scale and ordered series of numbers formation module comprise:
Pixel sampling unit, in units of the pixel in each region, carries out the pixel gradation intervals sampling setting step-length;
Gray scale and computing unit, calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
Wherein, described setting step-length can be 5 pixels.
Preferably, also comprise:
Module chosen by gray scale and ordered series of numbers sample, for multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version to bank note to be identified;
Network parameter training module, for adopting BP neural network to described sample training, obtains the network parameter of the recognizer of BP neural network.
Wherein, described first setting quantity can be 20.
The said goods can perform the method that any embodiment of the present invention provides, and possesses the corresponding functional module of manner of execution and beneficial effect.
Note, above are only preferred embodiment of the present invention and institute's application technology principle.Skilled person in the art will appreciate that and the invention is not restricted to specific embodiment described here, various obvious change can be carried out for a person skilled in the art, readjust and substitute and can not protection scope of the present invention be departed from.Therefore, although be described in further detail invention has been by above embodiment, the present invention is not limited only to above embodiment, when not departing from the present invention's design, can also comprise other Equivalent embodiments more, and scope of the present invention is determined by appended right.
Claims (10)
1. bank note is towards a recognition methods, it is characterized in that, described method comprises:
Obtain the banknote image of bank note to be identified;
According to setting division rule, described banknote image is divided into the region of the first setting quantity;
Calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity;
The recognizer of described ordered series of numbers input BP neural network is identified, with identify obtain bank note to be identified towards, wherein, the network parameter of the recognizer of described BP neural network is obtained by training.
2. method according to claim 1, is characterized in that, before by the recognizer of described ordered series of numbers input BP neural network, also comprises:
Described ordered series of numbers is normalized, obtains normalization ordered series of numbers.
3. method according to claim 1, is characterized in that, according to setting division rule, the region described banknote image being divided into the first setting quantity comprises:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
4. method according to claim 1, is characterized in that, calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity, comprising:
In units of pixel in each region, carry out the pixel gradation intervals sampling setting step-length;
Calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
5., according to the arbitrary described method of claim 1-4, it is characterized in that, also comprise:
To multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version of bank note to be identified;
Adopt BP neural network to described sample training, obtain the network parameter of the recognizer of BP neural network.
6. bank note is towards a recognition device, it is characterized in that, described device comprises:
Banknote image acquisition module, for obtaining the banknote image of bank note to be identified;
Image-region divides module, for according to setting division rule, described banknote image is divided into the region of the first setting quantity;
Gray scale and ordered series of numbers form module, for calculate each region in described banknote image gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity;
Bank note towards identification module, for the recognizer of described ordered series of numbers input BP neural network is identified, with identify obtain bank note to be identified towards, wherein, the network parameter of the recognizer of described BP neural network is obtained by training.
7. device according to claim 6, is characterized in that, also comprises:
Ordered series of numbers normalization module, for before the recognizer described ordered series of numbers being inputted BP neural network, is normalized described ordered series of numbers, obtains normalization ordered series of numbers.
8. device according to claim 6, is characterized in that, described image-region divide module specifically for:
According to matrix form, described banknote image is divided into the first setting quantity etc. subregion.
9. device according to claim 6, is characterized in that, described gray scale and ordered series of numbers form module and comprise:
Pixel sampling unit, in units of the pixel in each region, carries out the pixel gradation intervals sampling setting step-length;
Gray scale and computing unit, calculate sampling after each region gray scale and, form gray scale and ordered series of numbers that one has the first setting quantity.
10., according to the arbitrary described device of claim 6-9, it is characterized in that, also comprise:
Module chosen by gray scale and ordered series of numbers sample, for multiple gray scale and ordered series of numbers samples towards choosing the second setting quantity respectively of the value of money version to bank note to be identified;
Network parameter training module, for adopting BP neural network to described sample training, obtains the network parameter of the recognizer of BP neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510250574.9A CN104851183A (en) | 2015-05-15 | 2015-05-15 | Paper currency face and orientation recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510250574.9A CN104851183A (en) | 2015-05-15 | 2015-05-15 | Paper currency face and orientation recognition method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104851183A true CN104851183A (en) | 2015-08-19 |
Family
ID=53850805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510250574.9A Pending CN104851183A (en) | 2015-05-15 | 2015-05-15 | Paper currency face and orientation recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104851183A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106408747A (en) * | 2016-09-06 | 2017-02-15 | 深圳怡化电脑股份有限公司 | Image recognition method and apparatus |
| CN106485828A (en) * | 2016-09-18 | 2017-03-08 | 深圳怡化电脑股份有限公司 | A kind of Paper Currency Identification and device |
| CN106548202A (en) * | 2016-10-25 | 2017-03-29 | 深圳怡化电脑股份有限公司 | Note denomination recognition methodss and system |
| CN107680248A (en) * | 2017-10-13 | 2018-02-09 | 深圳怡化电脑股份有限公司 | A kind of bank note towards recognition methods, device, ATM and storage medium |
| CN107784730A (en) * | 2017-09-14 | 2018-03-09 | 深圳怡化电脑股份有限公司 | A kind of recognition methods of bank note and device |
| CN108320372A (en) * | 2018-01-22 | 2018-07-24 | 中南大学 | A kind of folding Paper Currency Identification |
| CN113936374A (en) * | 2021-09-26 | 2022-01-14 | 中国农业银行股份有限公司四川省分行 | Paper currency identification method based on double-attention machine system |
-
2015
- 2015-05-15 CN CN201510250574.9A patent/CN104851183A/en active Pending
Non-Patent Citations (2)
| Title |
|---|
| 崔明星 唐降龙 沙宗先: "一种基于神经网络的纸币识别算法", 《中国智能自动化会议》 * |
| 郭玉佳: "纸币图像处理技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106408747A (en) * | 2016-09-06 | 2017-02-15 | 深圳怡化电脑股份有限公司 | Image recognition method and apparatus |
| CN106485828A (en) * | 2016-09-18 | 2017-03-08 | 深圳怡化电脑股份有限公司 | A kind of Paper Currency Identification and device |
| CN106485828B (en) * | 2016-09-18 | 2019-04-26 | 深圳怡化电脑股份有限公司 | A kind of Paper Currency Identification and device |
| CN106548202A (en) * | 2016-10-25 | 2017-03-29 | 深圳怡化电脑股份有限公司 | Note denomination recognition methodss and system |
| CN107784730A (en) * | 2017-09-14 | 2018-03-09 | 深圳怡化电脑股份有限公司 | A kind of recognition methods of bank note and device |
| CN107680248A (en) * | 2017-10-13 | 2018-02-09 | 深圳怡化电脑股份有限公司 | A kind of bank note towards recognition methods, device, ATM and storage medium |
| CN107680248B (en) * | 2017-10-13 | 2019-09-20 | 深圳怡化电脑股份有限公司 | A kind of bank note towards recognition methods and device |
| CN108320372A (en) * | 2018-01-22 | 2018-07-24 | 中南大学 | A kind of folding Paper Currency Identification |
| CN113936374A (en) * | 2021-09-26 | 2022-01-14 | 中国农业银行股份有限公司四川省分行 | Paper currency identification method based on double-attention machine system |
| CN113936374B (en) * | 2021-09-26 | 2024-03-15 | 中国农业银行股份有限公司四川省分行 | Paper currency identification method based on double-attention mechanism |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104851183A (en) | Paper currency face and orientation recognition method and device | |
| CN109685115B (en) | Fine-grained conceptual model with bilinear feature fusion and learning method | |
| CN110532920B (en) | Face recognition method for small-quantity data set based on FaceNet method | |
| CN107273784B (en) | Image pattern recognition apparatus and method | |
| CN102081734B (en) | Object detecting device and its learning device | |
| CN108830188A (en) | Vehicle checking method based on deep learning | |
| CN109002755B (en) | Age estimation model construction method and estimation method based on face image | |
| WO2016037300A1 (en) | Method and system for multi-class object detection | |
| CN107122375A (en) | The recognition methods of image subject based on characteristics of image | |
| CN104866810A (en) | Face recognition method of deep convolutional neural network | |
| CN104484658A (en) | Face gender recognition method and device based on multi-channel convolution neural network | |
| CN110309835B (en) | Image local feature extraction method and device | |
| CN110222718B (en) | Image processing method and device | |
| CN110222607B (en) | Method, device and system for detecting key points of human face | |
| CN110490265B (en) | Image steganalysis method based on double-path convolution and feature fusion | |
| CN111046917B (en) | Object-based enhanced target detection method based on deep neural network | |
| CN104794455B (en) | A kind of Dongba pictograph recognition methods | |
| CN111414875B (en) | Three-dimensional point cloud head posture estimation system based on depth regression forest | |
| CN110363238A (en) | Intelligent vehicle damage identification method, system, electronic equipment and storage medium | |
| CN114140665A (en) | Dense small target detection method based on improved YOLOv5 | |
| CN114492634A (en) | Fine-grained equipment image classification and identification method and system | |
| Gawade et al. | Early-stage apple leaf disease prediction using deep learning | |
| CN113313688B (en) | Energetic material medicine barrel identification method and system, electronic equipment and storage medium | |
| CN111275070B (en) | Signature verification method and device based on local feature matching | |
| CN109948577B (en) | Cloth identification method and device and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150819 |
|
| RJ01 | Rejection of invention patent application after publication |