CN109409227A - A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN - Google Patents

A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN Download PDF

Info

Publication number
CN109409227A
CN109409227A CN201811113707.8A CN201811113707A CN109409227A CN 109409227 A CN109409227 A CN 109409227A CN 201811113707 A CN201811113707 A CN 201811113707A CN 109409227 A CN109409227 A CN 109409227A
Authority
CN
China
Prior art keywords
quality
feature vector
cnn
image
layer
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
Application number
CN201811113707.8A
Other languages
Chinese (zh)
Inventor
秦传波
谌瑶
曾军英
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuyi University Fujian
Original Assignee
Wuyi University Fujian
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 Wuyi University Fujian filed Critical Wuyi University Fujian
Priority to CN201811113707.8A priority Critical patent/CN109409227A/en
Publication of CN109409227A publication Critical patent/CN109409227A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/12—Fingerprints or palmprints
    • G06V40/1347—Preprocessing; Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Human Computer Interaction (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a kind of finger vena plot quality appraisal procedures and its device based on multichannel CNN.The gray level image of input is subjected to binarization operation and obtains binary image, gray level image and binary image are sent in corresponding CNN network simultaneously and carry out feature extraction, it the feature vector that obtains will be extracted carries out fusion and obtain fusion feature vector, it is classified to preset mass candidates class after calculating by full articulamentum fusion feature vector, the result according to affiliated candidate class as assessment.Multi-channel synchronous feature extraction is realized, quick and precisely the quality of finger vena figure is assessed.

Description

A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN
Technical field
The present invention relates to field of biological recognition, especially a kind of finger vena plot quality assessment side based on multichannel CNN Method and its device.
Background technique
Currently, finger vena identification is as emerging biometrics identification technology, it is the important composition of the following identification technology Part.And in collection process, since acquisition equipment and the situation for being collected individual are different, the finger vena acquired every time It is low that the quality of figure also has height to have, and low-quality finger vena figure is usually relatively fuzzyyer, if by low-quality finger vena caption volume Into database, it will seriously affect feature extraction and matching when later period application, cause the recognition performance of system poor, so needing Low-quality finger vena figure is rejected in collection process, the finger vena figure of high quality is registered in database, therefore How to carry out assessment to the quality of finger vena figure is committed step.
It is substantially the figure using computer technology to finger vena figure for the quality evaluation of finger vena figure Feature extracts, and can be divided into the Feature Parameter Fusion side of manual designs in conventional method according to the difference of feature extraction mode Method, vein point number statistical method.Although these methods, which can lead to, realizes quality evaluation, often very when encountering complicated image Hardly possible by manual designs go out effectively and robust feature extracting method, vein point number statistical method need to finger vena figure into A large amount of image procossing early period of row, process is complicated and takes a long time.It is most of using individually for binary picture in the prior art Picture or grayscale image are identified, but grayscale image generally comprises more noise, and binary image is easy during binaryzation Information is lost, therefore the characteristic pattern accuracy obtained is lower.
Summary of the invention
It can be right by CNN (convolutional neural networks) to solve the above problems, the purpose of the present invention is to provide a kind of methods The two-value of finger vena figure is merged after drawing the feature extraction with grayscale image, carries out matter to finger vena figure using fusion feature Amount assessment.Multichannel extracts, and improves the accuracy of characteristic pattern.
Technical solution used by the present invention solves the problems, such as it is: a kind of finger vena plot quality based on multichannel CNN Appraisal procedure, comprising the following steps:
The gray level image for reading input carries out binarization operation to the gray level image, obtains binary image；
The binary image and gray level image are respectively sent to progress feature in the convolutional layer of corresponding CNN network to mention It takes, obtains binaryzation feature vector and gray feature vector；
By the binaryzation feature vector and gray feature vector fused in tandem, fusion feature vector is obtained, and be sent to In the full articulamentum of CNN network, class probability vector is obtained by flexible max calculation；
Preset mass candidates class is read, the class probability vector is classified in corresponding mass candidates class, is completed Quality evaluation.
Further, it is described to the gray level image carry out binarization operation include: to gray level image execute gray scale normalization, Based on direction paddy shape detection enhancing and local adaptive threshold fuzziness.
Further, the convolutional layer of the CNN network includes gray scale training convolutional layer and binaryzation training convolutional layer；It is described The convolutional layer of CNN network includes 4 layers of convolutional network, and the convolutional network includes convolutional layer, pond layer and random deactivating layer.
Further, the convolution kernel of the convolutional layer is 3 × 3, and convolution step-length is 1；The template of the pond layer is 2 × 2, step A length of 2.
Further, the full articulamentum includes the first full articulamentum and the second full articulamentum, the first full articulamentum packet 512 neurons are included, the second full articulamentum includes 2 neurons.
Further, the mass candidates class includes high quality class and low quality class.
A kind of finger vena plot quality assessment device based on multichannel CNN, including following device:
Binary image acquisition device carries out binaryzation behaviour to the gray level image for reading the gray level image of input Make, obtains binary image；
Feature vector acquisition device, for the binary image and gray level image to be respectively sent to corresponding CNN net Feature extraction is carried out in the convolutional layer of network, obtains binaryzation feature vector and gray feature vector；
Fusion feature vector acquisition device is used for the binaryzation feature vector and gray feature vector fused in tandem, Obtain fusion feature vector, and be sent in the full articulamentum of CNN network, by flexible max calculation obtain class probability to Amount；
The class probability vector is classified to corresponding by quality assessment device for reading preset mass candidates class In mass candidates class, quality evaluation is completed.
Further, the binary image acquisition device further includes following device:
Binarization operation device, for executing gray scale normalization, based on the detection enhancing drawn game of direction paddy shape to gray level image Portion's adaptive threshold fuzziness.
A kind of finger vena plot quality assessment device based on multichannel CNN, comprising:
At least one processor；And
The memory being connect at least one described processor communication；
Wherein, the memory is stored with the instruction that can be executed by least one described processor, and described instruction is described At least one processor executes, so that at least one described processor is able to carry out one kind as described above and is based on multichannel The finger vena plot quality appraisal procedure of CNN.
A kind of non-transitorycomputer readable storage medium, the computer-readable recording medium storage have computer that can hold Row instruction, the computer executable instructions are for making computer execute a kind of hand based on multichannel CNN as described above Refer to vein plot quality appraisal procedure.
The beneficial effects of the present invention are: the present invention uses a kind of finger vena plot quality assessment side based on multichannel CNN Method and its device.Binary image is obtained according to the gray level image of input, and the two is respectively sent to corresponding CNN network simultaneously Feature extraction is carried out, then the feature vector obtained is subjected to fusion and obtains fusion feature vector, it is special to fusion using full articulamentum Vector Processing is levied, and then completes the classification of mass candidates class, the quality evaluation according to locating classification as input picture.It compares The method that the prior art only extracts gray level image or binary image is played, method of the invention can be effectively by gray level image and two-value Change image to be merged, solves the problems, such as that extracted picture noise is excessive in the prior art and information is lost, realize matter Measure the Accuracy and high efficiency of assessment.
Detailed description of the invention
The invention will be further described with example with reference to the accompanying drawing.
Fig. 1 is a kind of method flow diagram of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention；
Fig. 2 is a kind of volume of the first embodiment of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention Product schematic network structure；
Fig. 3 is a kind of CNN network diagram of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention；
Fig. 4 is a kind of detailed step figure of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention；
Fig. 5 is a kind of volume of the second embodiment of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention Product schematic network structure；
Fig. 6 is a kind of volume of the 3rd embodiment of the finger vena plot quality appraisal procedure based on multichannel CNN of the present invention Product schematic network structure.
Drawing reference numeral explanation:
1. binary image；2. gray level image；3. binaryzation convolution trains layer；4. gray scale convolution trains layer；5. full connection Layer.
Specific embodiment
Referring to Fig.1, a kind of finger vena plot quality appraisal procedure based on multichannel CNN of the invention, including following step It is rapid:
The gray level image for reading input carries out binarization operation to the gray level image, obtains binary image；
The binary image and gray level image are respectively sent to progress feature in the convolutional layer of corresponding CNN network to mention It takes, obtains binaryzation feature vector and gray feature vector；
By the binaryzation feature vector and gray feature vector fused in tandem, fusion feature vector is obtained, and be sent to In the full articulamentum of CNN network, class probability vector is obtained by flexible max calculation；
Preset mass candidates class is read, the class probability vector is classified in corresponding mass candidates class, is completed Quality evaluation.
Wherein, after carrying out binarization operation to the gray level image, also retain the gray level image for subsequent operation.
Preferably, the CNN network is preparatory trained depth convolutional neural networks.
Wherein, when classifying to the class probability vector, using biggish one of generic probability as selected Classification.
Further, it is described to the gray level image carry out binarization operation include: to gray level image execute gray scale normalization, Based on direction paddy shape detection enhancing and local adaptive threshold fuzziness.
Wherein, when detecting input gray level image, gray value of image is normalized to [0,255], normalization formula is
Wherein, carrying out enhancing by the way of based on the detection enhancing of direction paddy shape can preferably be partitioned into veinprint Come, while the operation based on the detection enhancing of direction paddy shape only carries out integer calculations, to compared with other increasings using floating type operation Time consumed by strong method is shorter, is conducive to accelerate recognition speed.
Wherein, the local auto-adaptive Threshold segmentation determines each picture according to the distribution of enhanced image peripheral pixel The segmentation threshold of element uses local auto-adaptive Threshold segmentation to enhanced image since the figure of finger vena figure is irregular Be conducive to the image for dividing brightness or contrast distribution unevenness.
Referring to figs. 2 and 3, further, the convolutional layer of the CNN network includes gray scale training convolutional layer 4 and binaryzation instruction Practice convolutional layer 3；The convolutional layer of the CNN network include 4 layers of convolutional network, the convolutional network include convolutional layer, pond layer and Random deactivating layer.
Wherein, every layer of convolutional network is by being all the linear function of upper layer output, therefore can exist between each parameter mutual Interdependent relationship, it is easy to cause hierarchy of 00operation more deeper more complicated.In order to improve the efficiency of operation, transported in each layer of convolutional network Activation primitive is introduced after having calculated to be activated, and is made the output 0 of partial nerve member, is thinned out network, reduces between parameter Relation of interdependence.
Preferably, preceding 3 layers of activation primitive uses Relu activation primitive, and the 4th layer uses softmax activation primitive.
Preferably, the random inactivation probability of the random deactivating layer of preceding level 2 volume product network is 0.3, rear level 2 volume product network with The random inactivation probability of machine deactivating layer is 0.5.
Further, the convolution kernel of the convolutional layer is 3 × 3, and convolution step-length is 1；The template of the pond layer is 2 × 2, step A length of 2.
Wherein, after 4 layers of convolutional network, the characteristic dimension of the binaryzation feature vector and gray feature vector is 512 dimensions, therefore the dimension of resulting fusion feature vector is 1024 dimensions.
Further, the full articulamentum includes the first full articulamentum and the second full articulamentum, the first full articulamentum packet 512 neurons are included, the second full articulamentum includes 2 neurons.
Wherein, it since the dimension of the fusion feature vector is 1024 dimensions, is tieed up by residue 512 after the first full articulamentum, It is tieed up by residue 2 after the second full articulamentum, by being classified as high quality class and low quality class to what image carried out in this present embodiment, Therefore the second full articulamentum selects 2 neurons to realize classification.
Further, the mass candidates class includes high quality class and low quality class.
Wherein, the classification standard of the mass candidates class is as obtained by training in advance.
With reference to Fig. 4, disposed of in its entirety step of the present invention is illustrated by way of example below:
Step 101, the gray level image for reading input carries out binarization operation to the gray level image, obtains binary picture Picture；
Step 102, gray level image is sent to the gray scale training convolutional layer of CNN network；Binary image is sent to CNN The binaryzation training convolutional layer of network；
Step 103, gray level image is trained in gray scale training convolutional layer, obtains the gray feature vector of 512 dimensions； Binary image is trained in binaryzation training convolutional layer, obtains the binaryzation feature vector of 512 dimensions；
Step 104, gray feature vector sum binaryzation feature vector is merged, obtains the fusion feature vector of 1024 dimensions；
Step 105, the fusion feature vector is sent to full articulamentum, by the flexible most matter of fundamental importance in full articulamentum It calculates, obtains class probability vector；
Step 106, preset mass candidates class is read, by class probability vector point into the highest candidate class of probability；
Step 107, if the class probability vector is low quality classification, low quality hand is set by the grayscale image of input Refer to vein figure；If the class probability vector is high quality classification, set the grayscale image of input to do quality finger vena Figure.
A kind of finger vena plot quality assessment device based on multichannel CNN, including following device:
Binary image acquisition device carries out binaryzation behaviour to the gray level image for reading the gray level image of input Make, obtains binary image；
Feature vector acquisition device, for the binary image and gray level image to be respectively sent to corresponding CNN net Feature extraction is carried out in the convolutional layer of network, obtains binaryzation feature vector and gray feature vector；
Fusion feature vector acquisition device is used for the binaryzation feature vector and gray feature vector fused in tandem, Obtain fusion feature vector, and be sent in the full articulamentum of CNN network, by flexible max calculation obtain class probability to Amount；
The class probability vector is classified to corresponding by quality assessment device for reading preset mass candidates class In mass candidates class, quality evaluation is completed.
Further, the binary image acquisition device further includes following device:
Binarization operation device, for executing gray scale normalization, based on the detection enhancing drawn game of direction paddy shape to gray level image Portion's adaptive threshold fuzziness.
A kind of finger vena plot quality assessment device based on multichannel CNN, comprising:
At least one processor；And
The memory being connect at least one described processor communication；
Wherein, the memory is stored with the instruction that can be executed by least one described processor, and described instruction is described At least one processor executes, so that at least one described processor is able to carry out one kind as described above and is based on multichannel The finger vena plot quality appraisal procedure of CNN.
A kind of non-transitorycomputer readable storage medium, the computer-readable recording medium storage have computer that can hold Row instruction, the computer executable instructions are for making computer execute a kind of hand based on multichannel CNN as described above Refer to vein plot quality appraisal procedure.
With reference to Fig. 5, a kind of finger vena plot quality appraisal procedure based on multichannel CNN, basic procedure and first is in fact Apply that example is essentially identical, there is following difference: the convolutional layer of the CNN network includes 3 layers of convolutional network, wherein first 2 layers of activation letter Number uses Relu activation primitive, and the 3rd layer uses softmax activation primitive；The random mistake of the random deactivating layer of preceding level 2 volume product network Probability living is that the random inactivation probability of the random deactivating layer of the 0.3, the 3rd layer of convolutional network is 0.5.
With reference to Fig. 6, a kind of finger vena plot quality appraisal procedure based on multichannel CNN, basic procedure and first is in fact Apply that example is essentially identical, there is following difference: the convolutional layer of the CNN network includes 5 layers of convolutional network, wherein first 4 layers of activation letter Number uses Relu activation primitive, and the 5th layer uses softmax activation primitive；The random mistake of the random deactivating layer of preceding level 2 volume product network Probability living is 0.3, and the random inactivation probability of the random deactivating layer of rear 3 layers of convolutional network is 0.5.
Through the above description of the embodiments, those of ordinary skill in the art can be understood that each embodiment The mode of general hardware platform can be added to realize by software, naturally it is also possible to pass through hardware.Those of ordinary skill in the art can With understand all or part of the process realized in above-described embodiment method be can be instructed by computer program it is relevant hard Part is completed, and the program can be stored in a computer-readable storage medium, the program is when being executed, it may include as above State the process of the embodiment of each method.Wherein, the storage medium can be magnetic disk, CD, read-only memory (Read- Only Memory, ROM) or random access memory (Random Access Memory, RAM) etc..
The above, only presently preferred embodiments of the present invention, the invention is not limited to above embodiment, as long as It reaches technical effect of the invention with identical means, all should belong to protection scope of the present invention.

Claims (10)

1.一种基于多通道CNN的手指静脉图质量评估方法，其特征在于，包括以下步骤：1. a finger vein map quality assessment method based on multi-channel CNN, is characterized in that, comprises the following steps: 读取输入的灰度图像，对所述灰度图像进行二值化操作，得出二值化图像；Read the input grayscale image, perform a binarization operation on the grayscale image, and obtain a binarized image; 将所述二值化图像和灰度图像分别发送至对应的CNN网络的卷积层中进行特征提取，得出二值化特征向量和灰度特征向量；The binarized image and the grayscale image are respectively sent to the corresponding convolutional layers of the CNN network for feature extraction, and the binarized feature vector and the grayscale feature vector are obtained; 将所述二值化特征向量和灰度特征向量串联融合，得出融合特征向量,并发送至CNN网络的全连接层中，通过柔性最大计算得出分类概率向量；The binarized feature vector and the grayscale feature vector are fused in series to obtain a fusion feature vector, which is sent to the fully connected layer of the CNN network, and a classification probability vector is obtained through maximum flexibility calculation; 读取预设的质量候选类，将所述分类概率向量分类至对应的质量候选类中，完成质量评估。Read the preset quality candidate classes, classify the classification probability vector into the corresponding quality candidate classes, and complete the quality assessment. 2.根据权利要求1所述的一种基于多通道CNN的手指静脉图质量评估方法，其特征在于，所述对所述灰度图像进行二值化操作包括：对灰度图像执行灰度归一化、基于方向谷形检测增强和局部自适应阈值分割。2. The method for evaluating the quality of finger vein images based on multi-channel CNN according to claim 1, wherein the performing a binarization operation on the grayscale image comprises: performing grayscale normalization on the grayscale image. Unification, directional valley detection enhancement and locally adaptive threshold segmentation. 3.根据权利要求1所述的一种基于多通道CNN的手指静脉图质量评估方法，其特征在于：所述CNN网络的卷积层包括灰度训练卷积层和二值化训练卷积层；所述CNN网络的卷积层包括4层卷积网络，所述卷积网络包括卷积层、池化层和随机失活层。3. a kind of finger vein map quality assessment method based on multi-channel CNN according to claim 1, is characterized in that: the convolution layer of described CNN network comprises gray level training convolution layer and binarization training convolution layer ; The convolutional layer of the CNN network includes a 4-layer convolutional network, and the convolutional network includes a convolutional layer, a pooling layer and a random deactivation layer. 4.根据权利要求3所述的一种基于多通道CNN的手指静脉图质量评估方法，其特征在于：所述卷积层的卷积核为3×3，卷积步长为1；所述池化层的模板为2×2，步长为2。4. A method for evaluating finger vein image quality based on multi-channel CNN according to claim 3, wherein the convolution kernel of the convolution layer is 3×3, and the convolution step size is 1; the The template of the pooling layer is 2×2 and the stride is 2. 5.根据权利要求1所述的一种基于多通道CNN的手指静脉图质量评估方法，其特征在于：所述全连接层包括第一全连接层和第二全连接层，所述第一全连接层包括512个神经元，所述第二全连接层包括2个神经元。5. A method for evaluating finger vein map quality based on multi-channel CNN according to claim 1, wherein the fully connected layer comprises a first fully connected layer and a second fully connected layer, and the first fully connected layer The connected layer includes 512 neurons, and the second fully connected layer includes 2 neurons. 6.根据权利要求1所述的一种基于多通道CNN的手指静脉图质量评估方法，其特征在于：所述质量候选类包括高质量类和低质量类。6 . The method for evaluating the quality of finger vein images based on a multi-channel CNN according to claim 1 , wherein the quality candidate classes include high-quality classes and low-quality classes. 7 . 7.一种基于多通道CNN的手指静脉图质量评估装置，其特征在于，包括以下装置：7. A device for evaluating the quality of finger vein diagrams based on multi-channel CNN, characterized in that, comprising the following devices: 二值化图像获取装置，用于读取输入的灰度图像，对所述灰度图像进行二值化操作，得出二值化图像；The binarized image acquisition device is used to read the input grayscale image, and perform a binarization operation on the grayscale image to obtain a binarized image; 特征向量获取装置，用于将所述二值化图像和灰度图像分别发送至对应的CNN网络的卷积层中进行特征提取，得出二值化特征向量和灰度特征向量；A feature vector acquisition device, for sending the binarized image and the grayscale image to the convolutional layer of the corresponding CNN network respectively for feature extraction, and obtaining a binarized feature vector and a grayscale feature vector; 融合特征向量获取装置，用于将所述二值化特征向量和灰度特征向量串联融合，得出融合特征向量,并发送至CNN网络的全连接层中，通过柔性最大计算得出分类概率向量；The fusion feature vector acquisition device is used to fuse the binarized feature vector and the grayscale feature vector in series to obtain the fusion feature vector, and send it to the fully connected layer of the CNN network, and obtain the classification probability vector through the maximum flexibility calculation ; 质量评估装置，用于读取预设的质量候选类，将所述分类概率向量分类至对应的质量候选类中，完成质量评估。The quality evaluation device is configured to read preset quality candidate classes, classify the classification probability vector into corresponding quality candidate classes, and complete the quality evaluation. 8.根据权利要求7所述的一种基于多通道CNN的手指静脉图质量评估装置，其特征在于，所述二值化图像获取装置还包括以下装置：二值化操作装置，用于对灰度图像执行灰度归一化、基于方向谷形检测增强和局部自适应阈值分割。8. A multi-channel CNN-based finger vein image quality assessment device according to claim 7, wherein the binarized image acquisition device further comprises the following device: a binarized operation device, used for graying The degree image performs grayscale normalization, directional valley-based detection enhancement, and locally adaptive threshold segmentation. 9.一种基于多通道CNN的手指静脉图质量评估装置，其特征在于，包括：9. A finger vein map quality assessment device based on multi-channel CNN is characterized in that, comprising: 至少一个处理器；以及，at least one processor; and, 与所述至少一个处理器通信连接的存储器；a memory communicatively coupled to the at least one processor; 其中，所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行如权利要求1-6任一项所述的方法。Wherein, the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute any one of claims 1-6 the method described. 10.一种非临时性计算机可读存储介质，其特征在于：所述计算机可读存储介质存储有计算机可执行指令，所述计算机可执行指令用于使计算机执行如权利要求1-6任一项所述的方法。10. A non-transitory computer-readable storage medium, characterized in that: the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to cause a computer to execute any one of claims 1-6 method described in item.
CN201811113707.8A 2018-09-25 2018-09-25 A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN Pending CN109409227A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811113707.8A CN109409227A (en) 2018-09-25 2018-09-25 A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811113707.8A CN109409227A (en) 2018-09-25 2018-09-25 A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN

Publications (1)

Publication Number Publication Date
CN109409227A true CN109409227A (en) 2019-03-01

Family

ID=65465088

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811113707.8A Pending CN109409227A (en) 2018-09-25 2018-09-25 A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN

Country Status (1)

Country Link
CN (1) CN109409227A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110084238A (en) * 2019-04-09 2019-08-02 五邑大学 Finger vena image segmentation method, device and storage medium based on LadderNet network
CN110251137A (en) * 2019-06-05 2019-09-20 长沙湖湘医疗器械有限公司 A kind of sleep detection method for noninvasive ventilator and the ventilator using this method
CN110292379A (en) * 2019-07-02 2019-10-01 重庆大学 The recognizer of epilepsy non-invasive diagnosis based on convolutional neural networks
CN110738141A (en) * 2019-09-26 2020-01-31 五邑大学 vein identification method, device, equipment and storage medium
WO2020199498A1 (en) * 2019-04-03 2020-10-08 平安科技（深圳）有限公司 Palmar digital vein comparison method and device, computer apparatus, and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8872909B2 (en) * 2010-06-10 2014-10-28 The Hong Kong Polytechnic University Method and apparatus for personal identification using finger imaging
CN106326886A (en) * 2016-11-07 2017-01-11 重庆工商大学 Finger-vein image quality evaluation method and system based on convolutional neural network
CN106529468A (en) * 2016-11-07 2017-03-22 重庆工商大学 Finger vein identification method and system based on convolutional neural network
KR102160184B1 (en) * 2017-06-02 2020-09-28 동국대학교 산학협력단 Finger vein recognition device and recognition method using convolutional neural network

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8872909B2 (en) * 2010-06-10 2014-10-28 The Hong Kong Polytechnic University Method and apparatus for personal identification using finger imaging
CN106326886A (en) * 2016-11-07 2017-01-11 重庆工商大学 Finger-vein image quality evaluation method and system based on convolutional neural network
CN106529468A (en) * 2016-11-07 2017-03-22 重庆工商大学 Finger vein identification method and system based on convolutional neural network
KR102160184B1 (en) * 2017-06-02 2020-09-28 동국대학교 산학협력단 Finger vein recognition device and recognition method using convolutional neural network

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
胡晶晶: "手指经脉图像质量评估算法综述", 《企业技术开发》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020199498A1 (en) * 2019-04-03 2020-10-08 平安科技（深圳）有限公司 Palmar digital vein comparison method and device, computer apparatus, and storage medium
JP2021533493A (en) * 2019-04-03 2021-12-02 平安科技(深▲せん▼)有限公司Ping An Technology (Shenzhen) Co., Ltd. Finger vein matching method, device, computer equipment, and storage medium
US11893773B2 (en) 2019-04-03 2024-02-06 Ping An Technology (Shenzhen) Co., Ltd. Finger vein comparison method, computer equipment, and storage medium
CN110084238A (en) * 2019-04-09 2019-08-02 五邑大学 Finger vena image segmentation method, device and storage medium based on LadderNet network
CN110084238B (en) * 2019-04-09 2023-01-03 五邑大学 Finger vein image segmentation method and device based on LadderNet network and storage medium
CN110251137A (en) * 2019-06-05 2019-09-20 长沙湖湘医疗器械有限公司 A kind of sleep detection method for noninvasive ventilator and the ventilator using this method
CN110292379A (en) * 2019-07-02 2019-10-01 重庆大学 The recognizer of epilepsy non-invasive diagnosis based on convolutional neural networks
CN110738141A (en) * 2019-09-26 2020-01-31 五邑大学 vein identification method, device, equipment and storage medium
WO2021056974A1 (en) * 2019-09-26 2021-04-01 五邑大学 Vein recognition method and apparatus, device, and storage medium

Similar Documents

Publication Publication Date Title
CN111325764B (en) A method of fruit image contour recognition
CN111178197B (en) Instance Segmentation Method of Cohesive Pigs in Group Breeding Based on Mask R-CNN and Soft-NMS Fusion
CN109409227A (en) A kind of finger vena plot quality appraisal procedure and its device based on multichannel CNN
CN113435460B (en) A recognition method for bright crystal granular limestone images
CN104809481B (en) A kind of natural scene Method for text detection based on adaptive Color-based clustering
WO2018072233A1 (en) Method and system for vehicle tag detection and recognition based on selective search algorithm
CN103996018B (en) Face identification method based on 4DLBP
CN104077579B (en) Facial expression recognition method based on expert system
CN109961049A (en) A method for cigarette brand recognition in complex scenes
CN104598885B (en) The detection of word label and localization method in street view image
CN105335725A (en) Gait identification identity authentication method based on feature fusion
CN114693616B (en) A rice disease detection method, device and medium based on improved target detection model and convolutional neural network
CN113159045A (en) Verification code identification method combining image preprocessing and convolutional neural network
CN107967442A (en) A kind of finger vein identification method and system based on unsupervised learning and deep layer network
CN105718866A (en) Visual target detection and identification method
CN107766860A (en) Natural scene image Method for text detection based on concatenated convolutional neutral net
CN111008647B (en) Sample extraction and image classification method based on void convolution and residual linkage
CN106874905A (en) A kind of method of the natural scene text detection based on self study Color-based clustering
CN106446925A (en) Dolphin identity recognition method based on image processing
CN102136074B (en) Man-machine interface (MMI) based wood image texture analyzing and identifying method
CN103218833A (en) Edge-reinforced color space maximally stable extremal region detection method
CN110874835A (en) Crop leaf disease resistance identification method, system, electronic device and storage medium
CN118097233A (en) A data and knowledge-driven image classification method
CN104866850B (en) A kind of optimization method of text image binaryzation
CN113516193B (en) Image processing-based red date defect identification and classification method and device

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

Application publication date: 20190301

RJ01 Rejection of invention patent application after publication