KR101730964B1 - Method and apparatus for cognizing coin by using binary pattern - Google Patents

Method and apparatus for cognizing coin by using binary pattern Download PDF

Info

Publication number
KR101730964B1
KR101730964B1 KR1020150077219A KR20150077219A KR101730964B1 KR 101730964 B1 KR101730964 B1 KR 101730964B1 KR 1020150077219 A KR1020150077219 A KR 1020150077219A KR 20150077219 A KR20150077219 A KR 20150077219A KR 101730964 B1 KR101730964 B1 KR 101730964B1
Authority
KR
South Korea
Prior art keywords
coin
pattern
image
binarization
binarization pattern
Prior art date
Application number
KR1020150077219A
Other languages
Korean (ko)
Other versions
KR20160141489A (en
Inventor
노용만
이승호
김형일
김세민
Original Assignee
한국과학기술원
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 한국과학기술원 filed Critical 한국과학기술원
Priority to KR1020150077219A priority Critical patent/KR101730964B1/en
Publication of KR20160141489A publication Critical patent/KR20160141489A/en
Application granted granted Critical
Publication of KR101730964B1 publication Critical patent/KR101730964B1/en

Links

Images

Classifications

    • G—PHYSICS
    • G07—CHECKING-DEVICES
    • G07D—HANDLING OF COINS OR VALUABLE PAPERS, e.g. TESTING, SORTING BY DENOMINATIONS, COUNTING, DISPENSING, CHANGING OR DEPOSITING
    • G07D5/00—Testing specially adapted to determine the identity or genuineness of coins, e.g. for segregating coins which are unacceptable or alien to a currency
    • G07D5/005—Testing the surface pattern, e.g. relief

Abstract

A method for recognizing a coin using a binary pattern according to the present invention includes the steps of acquiring a quadrangle image of a segmented quadrangle corresponding to a coin region, generating a differential image from the obtained coin image, Extracting an area binarization pattern by extracting an edge intensity variation pattern for divided subareas in a concentric ring and comparing the extracted binarization pattern with a reference binarization pattern of the training binarization pattern and training coin image, It may include a process of classifying the type of coin through measurement.

Description

TECHNICAL FIELD [0001] The present invention relates to a coin recognition method and a coin recognition method using a binarization pattern,

The present invention relates to a coin recognition method, and more particularly, to a coin recognition method and apparatus using a binarization pattern suitable for automatically recognizing and classifying coin types by applying an area binarization pattern technique to coin images will be.

In general, coin recognition is a widely used technology in the fields of financial institutions such as banks, markets, vending machines, and entertainment devices.

Conventional conventional coin recognition methods use information such as the weight, thickness, radius, etc. of coins. Such conventional coin recognition methods use fake coins, which make the weight, thickness, radius, There is a problem that can not be clearly distinguished.

[Conventional Method 1] is to recognize the type of coin by using the weight, radius, and thickness information of the coin. [Conventional Method 2] uses coin size and coin metal color information to identify coin The type of which is known.

Thus, there is a problem in that it can not clearly distinguish such conventional methods 1 and 2, or false coins having similar characteristics to actual coins.

On the other hand, image-based coin recognition is useful for solving the problems of the conventional methods described above. [Conventional method 3] estimates and aligns the directions of coins and then localizes the local contrast in the coin area And recognizes the type of coin by extracting features.

Conventional method 3 also has a problem that structural information of coins can not be used for recognition, and because it is based on pixel information of images, there is a fundamental problem that it is difficult to recognize corrupted coins.

[Conventional Method 1]

R. Huber, H. Ramoser, K. Mayer, H. Penz, and M. Rubik, "Classification of coins using an eigenspace approach," Pattern Recognition Letters 26 (1), 2005.

[Conventional Method 2]

H. R. Al-Zoubi, "Efficient coin recognition using a statistical approach," In Electro / Information Technology (EIT), 2010.

[Conventional Method 3]

Z. Guo, L. Zhang, and D. Zhang, "Rotation invariant texture classification using LBP variance (LBPV) with global matching," Pattern Recognition, 43 (3), 2010.

Korean Published Patent Application No. 2009-0112958 (Published on October 29, 2009)

[1] R. Huber, H. Ramoser, K. Mayer, H. Penz, and M. Rubik, "Classification of coins using an eigenspace approach," Pattern Recognition Letters 26 (1), 2005. [2] H. R. Al-Zoubi, "Efficient coin recognition using a statistical approach," In Electro / Information Technology (EIT), 2010. [3] Z. Guo, L. Zhang, and D. Zhang, "Rotation invariant texture classification using LBP variance (LBPV) with global matching," Pattern Recognition, 43 (3), 2010. Doherty, Aiden R., and Alan F. Smeaton. "Automatically segmenting lifelog data into events." Image Analysis for Multimedia Interactive Services, 2008. WIAMIS'08. Ninth International Workshop on. IEEE, 2008.

The present invention proposes a new coin recognition method capable of automatically classifying the types of coins by means of a dissimilarity measurement technique between an area binarization pattern extracted by differential imaging of coin images and a binarization pattern of training coin images .

In addition, the present invention proposes an image-based coin recognition technique robust to changes in form existing in coin images (for example, rotation of coin, damage of coin, etc.).

The problems to be solved by the present invention are not limited to those mentioned above, and another problem to be solved by the present invention can be clearly understood by those skilled in the art from the following description will be.

According to one aspect of the present invention, there is provided a method for processing a coin, comprising the steps of: obtaining a quadrangle image of a segmented quadrangle corresponding to a coin area; generating a differential image from the obtained coin image; extracting a region binarization pattern by extracting an edge magnitude change pattern for divided sub-regions in a ring, and extracting a region binarization pattern from the extracted binarization pattern and a reference binarization pattern of the training coin image, and a method of classifying the types of coins through measurement of dissimilarity.

The coin image of the present invention may be a quadrangle image of a segmented quadrangle.

The region binarization pattern of the present invention can be extracted through mapping of an intra region binarization pattern and an inter-region binarization pattern.

The intra-area binarization pattern of the present invention can be calculated by comparing magnitudes of gradient strengths of the sub-areas and magnitudes of the average gradients.

The inter-area binarization pattern of the present invention can be calculated through a magnitude comparison between the gradient intensities of sub-areas within a certain ring and the gradient intensities of sub-areas within the outer ring.

The classification process of the present invention can classify the types of coins based on the distance calculation result between the pattern vector of the coin image and the pattern vector of the training coin image.

In the classification process of the present invention, a coin image having a relatively shortest distance among the calculated distances may be found and determined as a type label of a coin image.

The classifying process of the present invention can be identified as a false coin when the distance of the coin image having the shortest distance deviates from a predetermined reference threshold value.

According to another aspect of the present invention, there is provided an image processing apparatus comprising a coin image acquiring unit for acquiring a coin image of a quadrangle segmented corresponding to a coin region from an input image, a differential image generating unit for generating a differential image from the obtained coin image, A pattern extracting block for extracting a region binarization pattern by extracting an edge magnitude change pattern for the divided sub-regions within a ring of concentric circles from the differential image, A nonlinearity measuring unit for measuring dissimilarity between reference binarization patterns of training coin images stored in the reference pattern DB, and a nonlinearity measuring unit for classifying the types of coin based on the measurement results of the noninitial diagram, A coin recognition apparatus using a binarization pattern including a coin sorting section for discriminating a coin is provided.

The pattern extracting block of the present invention may further comprise an intra RBP calculator for calculating an intra region binarization pattern from the differential image, an inter RBP calculator for calculating an inter-region binarization pattern from the differential image, And an anti-aliasing pattern extracting unit for converting the inter-area binarization pattern into a rotation and flipping robust region binary pattern (RFR) robust against rotation and inversion to extract the area binarization pattern.

The intra-RBP calculator of the present invention can calculate the intra-region binarization pattern by comparing magnitude differences between the gradient strengths of the sub-regions and the average gradients.

The inter-RBP calculator of the present invention can calculate the inter-area binarization pattern by comparing magnitudes of gradient intensities of sub-areas within a certain ring and gradient intensities of sub-areas within the outer ring.

The non-iris diopter measurement unit of the present invention can measure the non-iris diopter by calculating the distance between the pattern vector of the coin image and the pattern vector of the training coin image.

The coin sorting unit of the present invention can find a coin image having a relatively shortest distance among the calculated distances and determine the coin image type label.

The coin sorting unit of the present invention can discriminate the coin image as a false coin when the distance of the coin image having the shortest distance deviates from a predetermined reference threshold value.

The present invention can automatically classify the types of coins into high-speed and high-speed by measuring dissimilarity between the binarization pattern extracted from differential image of the coin image and the binarization pattern of the coin image for training, And can reduce the burden on expensive equipment purchases by making it possible to use it in a non-high-end PC environment by reducing the weight of the algorithm using the image pattern.

Also, the present invention can automatically classify the types of coins into high-speed and high-speed through dissimilarity measurement between binarization patterns to effectively block the use and circulation of maliciously falsified coins.

1 is a block diagram of a coin recognition apparatus using a binary pattern according to an embodiment of the present invention.
2 is a flowchart illustrating a main process of recognizing a type of a coin using a binary pattern according to an embodiment of the present invention.
3 is a conceptual diagram for explaining a concept of calculating a intra RFR and an inter RFR by defining a concentric ring and a divided sub-region in a ring.
4 is a mapping example for explaining a concept of grouping the same RBPs into one RFR.
FIG. 5 is an exemplary view of the 13 kinds of RFRs.
6 is an exemplary diagram of a distance matrix for measuring a distance between a coin image pattern vector and a training coin image pattern vector.

First, the advantages and features of the present invention, and how to accomplish them, will be clarified with reference to the embodiments to be described in detail with reference to the accompanying drawings. While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, It will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

In the following description of the present invention, detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. It is to be understood that the following terms are defined in consideration of the functions of the present invention, and may be changed according to intentions or customs of a user, an operator, and the like. Therefore, the definition should be based on the technical idea described throughout this specification.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram of a coin recognition apparatus using a binarization pattern according to an embodiment of the present invention. The coin image acquiring unit 102, the differential image generating unit 104, the pattern extracting block 106, The pattern extracting block 106 may include an intra RBP calculating unit 1061, an inter-RBP calculating unit 1063, and a binarization unit 1063. The intra- A pattern extracting unit 1065, and the like.

1, the coin image acquiring unit 102 acquires a quadrangle coin image segmented corresponding to a coin region from an externally input image, and delivers the same to the differential image generating unit 104 .

The derivative image generating unit 104 may provide a function of generating a derivative image from the quadrangle coin image transferred from the coin image acquiring unit 102 and then transferring the differential image to the pattern extracting block 106.

Assuming that the coin image is I (x, y), for example, the differential image M (x, y) of the coin image can be generated as shown in Equation 1 below.

[Equation 1]

Figure 112016109061244-pat00014

In Equation (1), x and y are coordinates of the image. Here, the differential image shows the strength and direction of the edge, which is useful for extracting pattern information from coins that are dirty or have been damaged for a long time.

Next, the pattern extracting block 106 may include an intra RBP calculating section 1061, an inter-RBP calculating section 1063, and a binarization pattern extracting section 1065, and as an example, as shown in FIG. 3 , The differential image can be defined as a ring of concentric circles and subdivided subregions within the ring where n (1? N? N) is an index indicating the number of rings and s (1? S S) is an index indicating which sub-region is within a certain ring.

The binarization patterns extracted from the sub-regions represent information of change in gradient magnitude values occurring in the coin region. This can be achieved by calculating the difference of the overall average gradient intensities of the coin regions, where the overall average gradient M mu of the coin regions can be defined as: < EMI ID = 2.0 >

&Quot; (2) "

Figure 112015052546215-pat00002

In the above equation (2), m (n, s) is the gradient strength of the s-th sub-region of the n-th ring. The gradient magnitude difference between the s-th sub-region of the n-th ring and the entire coin region can be defined as Equation (3) below.

&Quot; (3) "

Figure 112015052546215-pat00003

Then, the intensity difference D mu of the average gradients can be calculated as shown in Equation (4) below.

&Quot; (4) "

Figure 112015052546215-pat00004

That is, a region binary pattern (RBP) can be extracted using the difference between the gradient intensity of the sub-region and the intensity difference of the average gradient. The region binarization pattern includes two types of intra-RBP and inter- inter RBP.

To this end, the intra RBP calculation unit 1061 calculates the intra RBP from the differential image delivered from the differential image generation unit 104 and then transmits the intra RBP to the binarization pattern extraction unit 1065. This intra RBP is expressed by the following equation 5 Can be calculated by comparing the magnitude of the gradient strength difference of the subregion with the magnitude difference of the average gradient, which represents the overall distribution of the edge.

&Quot; (5) "

Figure 112015052546215-pat00005

Next, the inter-RBP calculation unit 1063 calculates the inter-RBP from the differential image delivered from the differential image generation unit 104 and then transmits the inter-RBP to the binarization pattern extraction unit 1065. This inter-RBP is calculated by the following Equation 6 By comparing the magnitudes of the gradient intensities of the sub-areas within a particular ring with the magnitudes of the gradient intensities of the sub-areas within the outer ring, as shown in FIG.

&Quot; (6) "

Figure 112015052546215-pat00006

Therefore, in order to make the intra-RBP and the inter-RBP calculated by Equations 5 and 6 robust to the coin rotation, the binarization pattern extractor 1065 performs a rotation and flipping robust region binary pattern), thereby extracting an area binarization pattern for the input coin image, and delivering the extracted area binarization pattern to the non-unidirectional measurement unit 108. That is, the binarization pattern extracting unit 1065 extracts an area binarization pattern from the differential image by extracting an edge magnitude change pattern for divided subareas in a concentric ring.

For example, as shown in FIG. 4, in the case of the same RBP when rotation and inversion are performed, grouping is performed and mapping is performed with one RFR, thereby extracting an area binarization pattern for the input coin image.

Here, the RFR, as an example, can have 13 kinds, as shown in Fig. 5, and these have 13 different values (1, 2, ..., 13). 1 2-dimensional pattern vector X = {x 1, ...., x 2N-1 } to be used for coin classification by concatenating N and N number of intra RFR and inter- Can be obtained.

Next, the non-illumine degree measurement unit 108 measures the degree of coincidence between the area binarization pattern corresponding to the input coin image delivered from the binarization pattern extracting unit 1065 in the pattern extracting block 106 and the binarization pattern corresponding to the training coin It is possible to provide functions such as measuring the dissimilarity between the reference binarization patterns of the image and transmitting the results of the nonuniformity measurement to the coin sorting section 112. For this purpose, the reference pattern DB 110 stores reference binarization patterns for a plurality of coins that can be classified or discriminated, and these reference binarization patterns are added to the DB or added to the DB Can be deleted.

That is, the non-iris diopter measurement unit 108 calculates the distance dist (X q , X r ) between the pattern vector X q of the input coin image and the pattern vector X r of the training coin image to discriminate the type of coin Can be measured as shown in Equation (7) below.

&Quot; (7) "

Figure 112015052546215-pat00007

In the above Equation (2), DM can be defined as a distance matrix for measuring the distance in a short time, for example, as shown in Fig.

The coin sorting unit 112 classifies the type of the coin (discriminates the type) or determines whether or not the coin is maliciously falsified based on the result of the non-inference measurement transmitted from the non-inductance measurement unit 108 And so on.

For example, a coin image having a relatively shortest distance among the distances calculated using Equation (7) can be found to determine the type label of the coin image, and the distance of the coin image having the relatively shortest distance can be determined It will be possible to judge it as a false coin when it exceeds the reference threshold.

Here, when classified as a normal type of coin, for example, the amount of coins may be added to the previous counted amount, or the input path for storing the coins (sorting by type) may be changed, and when it is determined as a malicious false coin For example, an administrator or an operator may temporarily suspend the coin sorting operation so as to recognize it audibly and audibly, and then generate audible and visual alarms.

Next, a series of processes for classifying and discriminating the types of coins based on the binary pattern using the coin recognition apparatus according to the present embodiment having the above-described configuration will be described in detail.

2 is a flowchart illustrating a main process of recognizing a type of a coin using a binary pattern according to an embodiment of the present invention.

2, the coin image acquiring unit 102 acquires a quadrangle coin image segmented corresponding to the coin area from an externally input image (step 202), and the differential image generating unit 104 acquires the obtained quadrangle image (Step 204).

Next, in the intra-RBP calculation unit 1061 in the pattern extraction block 106, the intra-RBP is calculated from the differential image by comparing magnitude differences between the gradient strengths of the sub-regions and the average gradients as in Equation (5) (Step 206).

The inter-RBP calculator 1063 in the pattern extracting block 106 compares the gradient intensities of the sub-areas in the specific ring with the gradient intensities of the sub-areas in the outer ring, as shown in Equation (6) Lt; RTI ID = 0.0 > RBP < / RTI >

Then, the binarization pattern extracting unit 1065 extracts a region binarization pattern for the input coin image, that is, extracts a concentric ring from the differential image by transforming it into a rotation and flipping robust region binary pattern (RFR) An area binarization pattern is extracted by extracting an edge magnitude variation pattern for the divided sub-areas (step 210).

Again, the non-idosomaly measuring unit 108 measures the dissimilarity between the region binarization pattern corresponding to the extracted input coin image and the reference binarization pattern of the training coin image obtained by searching the reference pattern DB 110 212). That is, as shown in Equation (7), the non-inductance measurement unit 108 measures the distance dist (dist) between the pattern vector X q of the coin image input for coin sorting (type classification) and the pattern vector X r of the coin image for training (X q , X r ) is measured.

Finally, the coin sorting unit 112 classifies (identifies) the type of the corresponding coin based on the result of measurement of the unidirectional azimuth between the area binarization pattern corresponding to the input coin image and the reference binarization pattern, or identifies a malicious counterfeit coin (Step 214).

It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the following claims. It is easy to see that this is possible. That is, the embodiments disclosed in the present invention are not intended to limit the scope of the present invention but to limit the scope of the present invention.

Therefore, the scope of protection of the present invention should be construed in accordance with the following claims, and all technical ideas within the scope of equivalents should be interpreted as being included in the scope of the present invention.

Claims (15)

Acquiring a quadrangle image of a segmented segment corresponding to a coin area;
Generating a differential image from the obtained coin image,
Extracting a region binarization pattern from the generated differential image by extracting an edge magnitude change pattern for subdivided subareas in a concentric ring,
A process of sorting the types of coins through dissimilarity measurement between the extracted region binarization pattern and the reference binarization pattern of the training coin image
The method comprising the steps of:
Claim 2 has been abandoned due to the setting registration fee. The method according to claim 1,
Wherein,
Segmented quadrature coin images
Coin recognition method using binary pattern.
The method according to claim 1,
The region binarization pattern may be a sub-
Extracted through the mapping of the intra region binarization pattern and the inter-region binarization pattern
Coin recognition method using binary pattern.
The method of claim 3,
The intra-region binarization pattern may include:
Calculated by comparing the magnitude of the gradient strength of the subregion and the magnitude difference of the average gradient
Coin recognition method using binary pattern.
The method of claim 3,
The inter-area binarization pattern includes:
Is calculated through a magnitude comparison of the gradient intensities of the sub-areas within a particular ring and the gradient intensities of the sub-areas within the outer ring
Coin recognition method using binary pattern.
The method according to claim 1,
The classifying process includes:
And classifies the type of the coin based on the distance calculation result between the pattern vector of the coin image and the pattern vector of the training coin image
Coin recognition method using binary pattern.
The method according to claim 6,
The classifying process includes:
A coin image having a relatively shortest distance among the calculated distances is searched for and determined as a type label of a coin image
Coin recognition method using binary pattern.
8. The method of claim 7,
The classifying process includes:
When the distance of the coin image having the shortest distance deviates from a preset reference threshold value, it is determined as a false coin
Coin recognition method using binary pattern.
A coin image acquiring unit for acquiring a quadrangle coin image segmented corresponding to a coin region from an input image;
A differential image generating unit for generating a differential image from the obtained coin image,
A pattern extracting block for extracting an area binarization pattern from the generated differential image by extracting an edge magnitude change pattern for divided subareas in a concentric ring,
A nonlinearity measuring unit for measuring dissimilarity between the extracted area binarization pattern and a reference binarization pattern of a training coin image stored in the reference pattern DB,
A coin sorting unit for sorting the types of the coins based on the measurement result of the non-inference chart,
The coin recognizing device using the binarization pattern.
10. The method of claim 9,
Wherein the pattern extracting block comprises:
An intra RBP calculation unit for calculating an intra region binarization pattern from the differential image,
An inter-RBP calculator for calculating an inter-region binarization pattern from the differential image,
Extracting the region binarization pattern by converting the calculated intra-region binarization pattern and inter-region binarization pattern into rotation and flipping robust region binary pattern (RFR) robust against rotation and inversion,
The coin recognizing device using the binarization pattern.
11. The method of claim 10,
The intra-RBP calculator calculates,
The intra-area binarization pattern is calculated by comparing magnitude of the gradient intensity of the sub-area and magnitude difference of the average gradient
Coin recognition system using binarization pattern.
11. The method of claim 10,
The inter-RBP calculator includes:
The inter-area binarization pattern is calculated through a magnitude comparison between the gradient intensities of the sub-areas in the specific ring and the gradient intensities of the sub-areas in the outer ring thereof
Coin recognition system using binarization pattern.
10. The method of claim 9,
Wherein the non-
Calculating a distance between the pattern vector of the coin image and the pattern vector of the training coin image to measure the non-inference
Coin recognition system using binarization pattern.
14. The method of claim 13,
Wherein the coin sorting unit comprises:
A coin image having a relatively shortest distance among the calculated distances is searched for and determined as a type label of a coin image
Coin recognition system using binarization pattern.
15. The method of claim 14,
Wherein the coin sorting unit comprises:
When the distance of the coin image having the shortest distance deviates from a preset reference threshold value, it is determined as a false coin
Coin recognition system using binarization pattern.
KR1020150077219A 2015-06-01 2015-06-01 Method and apparatus for cognizing coin by using binary pattern KR101730964B1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020150077219A KR101730964B1 (en) 2015-06-01 2015-06-01 Method and apparatus for cognizing coin by using binary pattern

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020150077219A KR101730964B1 (en) 2015-06-01 2015-06-01 Method and apparatus for cognizing coin by using binary pattern

Publications (2)

Publication Number Publication Date
KR20160141489A KR20160141489A (en) 2016-12-09
KR101730964B1 true KR101730964B1 (en) 2017-04-27

Family

ID=57574517

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020150077219A KR101730964B1 (en) 2015-06-01 2015-06-01 Method and apparatus for cognizing coin by using binary pattern

Country Status (1)

Country Link
KR (1) KR101730964B1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109389615A (en) * 2018-09-29 2019-02-26 佳都新太科技股份有限公司 Coin discriminating method and processing terminal based on deep learning convolutional neural networks

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000163587A (en) 1998-11-30 2000-06-16 Hakko Automation Kk Coin checking device and method and recording medium
JP2002032812A (en) 2000-07-13 2002-01-31 Toyo Commun Equip Co Ltd Disc-shaped subject recognition system

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100966899B1 (en) 2008-04-25 2010-06-30 주식회사 동구전자 Coin cognition apparatus

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000163587A (en) 1998-11-30 2000-06-16 Hakko Automation Kk Coin checking device and method and recording medium
JP2002032812A (en) 2000-07-13 2002-01-31 Toyo Commun Equip Co Ltd Disc-shaped subject recognition system

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109389615A (en) * 2018-09-29 2019-02-26 佳都新太科技股份有限公司 Coin discriminating method and processing terminal based on deep learning convolutional neural networks
CN109389615B (en) * 2018-09-29 2021-05-28 佳都科技集团股份有限公司 Coin identification method based on deep learning convolutional neural network and processing terminal

Also Published As

Publication number Publication date
KR20160141489A (en) 2016-12-09

Similar Documents

Publication Publication Date Title
US10510152B2 (en) Systems, methods, and devices for image matching and object recognition in images using textures
US9741116B2 (en) System and method for boundary classification and automatic polyp detection
US9294665B2 (en) Feature extraction apparatus, feature extraction program, and image processing apparatus
CN107346409A (en) Pedestrian recognition methods and device again
CN108291876A (en) System and method for the authenticity for detecting product
CN105809655A (en) Vehicle checking method and system
CN106133752A (en) Eye gaze is followed the tracks of
Tajbakhsh et al. A classification-enhanced vote accumulation scheme for detecting colonic polyps
Yu et al. Scene text localization using edge analysis and feature pool
CN106056122B (en) A kind of image zone duplicating stickup altering detecting method based on KAZE characteristic point
Thalji et al. Iris Recognition using robust algorithm for eyelid, eyelash and shadow avoiding
Tajbakhsh et al. Automatic polyp detection from learned boundaries
Muhimmah et al. Automated cervical cell nuclei segmentation using morphological operation and watershed transformation
CN111191708A (en) Automatic sample key point marking method, device and system
JP2014186520A (en) Image processing apparatus, image processing method, and program
Muhimmah et al. Analysis of features to distinguish epithelial cells and inflammatory cells in Pap smear images
KR101730964B1 (en) Method and apparatus for cognizing coin by using binary pattern
Lee et al. Vehicle model recognition in video
Chaki An efficient two-stage Palmprint recognition using Frangi-filter and 2-component partition method
CN107481255B (en) Method and device for determining number of people
CN105825504B (en) A kind of image copy detection method based on MSCR provincial characteristics
CN114926635A (en) Method for segmenting target in multi-focus image combined with deep learning method
Ai et al. Geometry preserving active polygon-incorporated sign detection algorithm
CN110222666B (en) Signature authentication method and system
Creusen et al. A semi-automatic traffic sign detection, classification, and positioning system

Legal Events

Date Code Title Description
A201 Request for examination
E902 Notification of reason for refusal
E701 Decision to grant or registration of patent right
GRNT Written decision to grant