CN112149521A - Palm print ROI extraction and enhancement method based on multitask convolutional neural network - Google Patents
Palm print ROI extraction and enhancement method based on multitask convolutional neural network Download PDFInfo
- Publication number
- CN112149521A CN112149521A CN202010916060.3A CN202010916060A CN112149521A CN 112149521 A CN112149521 A CN 112149521A CN 202010916060 A CN202010916060 A CN 202010916060A CN 112149521 A CN112149521 A CN 112149521A
- Authority
- CN
- China
- Prior art keywords
- palm print
- neural network
- convolution
- image
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013527 convolutional neural network Methods 0.000 title claims abstract description 21
- 238000000605 extraction Methods 0.000 title claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 42
- 238000013528 artificial neural network Methods 0.000 claims abstract description 27
- 238000012545 processing Methods 0.000 claims abstract description 11
- 230000002708 enhancing effect Effects 0.000 claims abstract description 7
- 238000001914 filtration Methods 0.000 claims abstract description 7
- 238000012795 verification Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 33
- 230000011218 segmentation Effects 0.000 claims description 10
- 238000003062 neural network model Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 238000003709 image segmentation Methods 0.000 claims description 6
- 238000011176 pooling Methods 0.000 claims description 6
- 238000012937 correction Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 4
- 239000000284 extract Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 210000004556 brain Anatomy 0.000 description 5
- 241000282414 Homo sapiens Species 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000037303 wrinkles Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 230000000366 juvenile effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 208000028280 polygenic inheritance Diseases 0.000 description 1
- 108090000623 proteins and genes Proteins 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 210000000697 sensory organ Anatomy 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
Images
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
- G06V40/107—Static hand or arm
- G06V40/11—Hand-related biometrics; Hand pose recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration by non-spatial domain filtering
-
- G06T5/70—
-
- 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
Abstract
A palm print ROI extraction and enhancement method based on a multitask convolution neural network comprises the following steps: 1) preparing a sample before training, namely firstly copying the sample into two parts, namely A and B, and carrying out BM3D denoising and Gabor wavelet filtering on the sample A so as to carry out image enhancement processing on the sample; marking the sample B, and respectively marking two interphalangeal valley points and a palm print ROI area of the training sample; finally, performing data expansion on the marked training sample; 2) training a multitask convolution neural network by using the training sample generated in the step 1) to obtain a network model for extracting and enhancing the palm print ROI; 3) and verifying the trained multi-task convolutional neural network model through the verification set, outputting a result and correcting the result. The method can realize extraction of the palm print interesting region with image enhancement from the general palm print with higher accuracy and robustness.
Description
Technical Field
The invention relates to the field Of palm print enhancement segmentation, in particular to a palm print ROI (region Of interest) extracting and enhancing method based on a multitask convolutional neural network.
Background
The existing biological feature recognition technology mainly comprises fingerprint recognition, face recognition, iris recognition, palm print recognition and the like. Because the shape of the palm print is determined by the gene of each person, even if the palm print is damaged in the future, the later grown prints can keep the same shape as the original prints, and therefore, the method is a biological identification method with great potential. The palm print mainly comprises three lines, which are respectively: papillary lines, wrinkles and flexor lines. The three main lines are inherent to human beings and have good stability. Although the papillary lines and wrinkles produce slight changes during juvenile growth, changes in the palmar lines will not be apparent after adulthood, and these changes are not produced in a short time and require a long period of time. The hong Kong university of Physician has conducted extensive research on human palmprints for up to four years, which concluded that the palmprints were characterized by stability. The palm prints have the characteristics of uniqueness, even if the palms of the same twins are different, the palm prints of the left hand and the right hand of the same person are also different. From the perspective of genetic inheritance, the palm print belongs to polygenic inheritance, has uniqueness as well as fingerprints, but has larger area than fingerprints, and can better show detailed characteristics.
Meanwhile, the neural network has strong self-learning capability and the function of quickly searching the optimal solution, and is rapidly developed in recent years. The inspiration of the neural network is derived from brain nerve cells of the human brain, the operation of the human brain does not directly obtain information from the retina, but obtains the rules of things through a complex layered structure of the brain by receiving stimulation signals through sensory organs, the definite layered hierarchical structure reduces the data volume, optimizes the processing efficiency, and deep learning is generated under the inspiration of the brain structure. This provides the possibility of enhancing and extracting the palm print through the neural network.
Disclosure of Invention
In order to overcome the defects of poor time complexity and robustness of the conventional palm print enhancement and ROI extraction, the invention provides a palm print ROI extraction and enhancement method based on a multitask convolutional neural network, and the palm print ROI extraction and enhancement method has higher robustness and reduces a time period by self-learning and quickly searching an optimal solution function of the neural network. The multi-task learning is a machine learning method for learning a plurality of related tasks together based on shared representation, and the plurality of tasks are generalized mutually, so that the learning effect is improved.
In order to achieve the purpose, the invention adopts the technical scheme that:
a palm print ROI extraction and enhancement method based on a multitask convolutional neural network comprises the following steps:
1) preparing a sample before training, namely firstly copying the sample into two parts, namely A and B; performing BM3D denoising and Gabor wavelet filtering on the sample A so as to perform image enhancement processing on the sample; marking the sample B, and respectively marking two interphalangeal valley points and a palm print ROI area of the training sample; finally, performing data expansion on the marked training sample;
2) training a multitask convolution neural network by using the training sample generated in the step 1) to obtain a network model for extracting and enhancing the palm print ROI;
3) and verifying the trained multi-task convolutional neural network model through the verification set, outputting a result and correcting the result.
Further, the step 1) comprises the following steps:
(11) the selected palm print images have various noises, and the purpose of denoising the images is to provide a denoised training sample for the subsequent neural network training and perform denoising operation on the palm print images by using a BM3D denoising algorithm;
(12) carrying out filtering enhancement operation on the de-noised palm print image by using Gabor wavelet, wherein Gabor wavelet transform is one of wavelet transform and is orthogonalization of Gabor transform, decomposing the palm print image by using the Gabor wavelet, then carrying out inverse transform, and screening important components by changing corresponding parameters in the process so as to achieve the purpose of image enhancement;
(13) marking the palm print of the training sample, wherein the marked position is a valley point between the middle finger and the other two fingers, and the marked size is x pixels;
(14) marking the palm print of the training sample, wherein the marking position is a palm print interesting region, and the marking size is y pixels x z pixels, such as y equals 144, z equals 128;
(15) and performing data expansion on the m palm print images after the operation, wherein the expansion mode is to perform random rotation and left-right turning on the images, and screen the images still retaining the main information of the palm prints to be used as training samples for neural network training.
Still further, the step 2) comprises the following steps:
(21) constructing a multitask convolution neural network model to enable the multitask convolution neural network model to simultaneously realize enhancement and segmentation of a palm print image, dividing the network into three parts, wherein the first part is a sharing part, the sharing part carries out down-sampling operation on the palm print image, inputs the image and carries out operation of 2 convolution modules, the size of a convolution kernel is 3 x 3, a characteristic layer is 64, and the output of the last convolution module is stored; performing maximum pooling operation on the result, performing convolution operation of 2 convolution modules, wherein the size of a convolution kernel is 3 x 3, a characteristic layer is 128, and the result is output and stored through the last convolution module; the subsequent processing is the same as the above, and the convolution is performed after twice pooling, the convolution kernel of the corresponding convolution module is 3 x 3, the characteristic layer is 256, the convolution kernel is 3 x 3, and the characteristic layer is 512;
(22) the second part is an independent part and is a palm print image segmentation part, the palm print image segmentation part is subjected to upsampling on the basis of the output of the first part, the upsampling process is 3 times of deconvolution and 3 times of cross operation of 2 convolution modules, and finally a palm print segmentation graph with the same size as the original image is output, the convolution kernel of the convolution layer of the part is 3 x 3, the sizes of the feature layers are 256, 28 and 64 in sequence, the convolution kernel of the last convolution operation is 1 x 1, and the feature layer is 1;
(23) the third part is an independent part and is a palm print image enhancement part, the palm print image enhancement part is subjected to upsampling on the basis of the output of the first part, the upsampling process is the cross operation of 3 deconvolution modules and 3 convolution modules, and finally a palm print segmentation graph with the same size as the original image is output, the convolution kernel of the convolution layer of the part is 3 x 3, the sizes of the characteristic layers are 256, 28 and 64 in sequence, the convolution kernel of the last convolution operation is 1 x 1, and the characteristic layer is 1;
(24) in the constructed multitask convolutional neural network, the activation functions of two shared convolutional layers are both ReLU, in the independent convolutional layers of the two tasks, the palm print ROI extracts the last layer of the network structure to be sigmoid, and the activation function of the palm print image enhancement network structure is ReLU;
(25) in the training process, the total loss function of the multitask convolutional neural network is as follows:
Loss1+2＝α*loss1+β*loss2
Loss1+2is the total loss function of the multi-task neural network, alpha and beta are the weight coefficients preset corresponding to each task, loss1、loss2Extracting a loss function and a palm print image loss function for a palm print ROI in the multitask convolution neural network respectively, wherein the process is as follows:
(251) loss function loss for palm print ROI extraction1Is defined as:
(252) loss function loss for palm print image enhancement2Is defined as:
wherein f isiRepresenting a predicted value; y isiRepresenting the true value.
Further, the step 3) comprises the following steps:
(31) inputting an image to be processed into a trained multitask convolution neural network, comparing an output result of the network with an original image in a test set, calculating a loss function, performing network back propagation until the loss function is stable, and outputting a result graph: a palm print calibration image and a palm print enhancement image;
(32) correcting the calibration graph of the region of interest of the palm print, wherein the correcting steps are as follows:
(321) connecting two valley points, drawing a straight line and a corresponding vertical line, establishing a new coordinate axis by using the two straight lines, taking the intersection point of the two straight lines as an original point, rotating the palm print image by taking the original point O as a center by an angle theta, and expressing the angle theta as follows:
when theta is a positive value, rotating the palm print image anticlockwise; otherwise, clockwise rotating the palm print image, and judging the distance between the palm print interested region automatically calibrated by the neural network and the valley point by taking the two valley points as reference objects;
(322) according to experience judgment, the distance between the left side boundary of the palm print interesting region and the valley point connecting line can be determined to be one fourth of the distance between the two valley points, the distance between the right side boundary is 1.5 times of the distance between the two valley points, if the condition is met, the calibration result does not need to be corrected, otherwise, correction processing needs to be carried out;
(323) if the palm print calibration result is not satisfied, automatically correcting the palm print calibration result, filling a small part of the palm print interesting region, and restoring the redundant part.
Compared with the prior art, the invention has the beneficial effects that: the palm print image enhancement method can be used for extracting the region of interest of the palm print, has an automatic correction function, adapts to palms of different sizes, and can also be used for enhancing the palm print image. The time cost of the respective operation of the two is reduced, and the robustness and the universality are better.
Drawings
FIG. 1 is a schematic diagram of palm print valley and ROI calibration in the present invention.
FIG. 2 is a schematic diagram of the structure of the multitask convolutional neural network in the present invention.
FIG. 3 is a flow chart of the multitask convolutional neural network training steps of the present invention.
Detailed Description
The invention will be further described with reference to the following figures and embodiments:
referring to fig. 1, 2 and 3, a palm print ROI extraction and enhancement method based on a multitask convolutional neural network includes the following steps:
1) preparing a sample before training, namely firstly copying the sample into two parts, namely A and B; performing BM3D denoising and Gabor wavelet filtering on the sample A so as to perform image enhancement processing on the sample; marking the sample B, and respectively marking two interphalangeal valley points and a palm print ROI area of the training sample; and finally, performing data expansion on the marked training sample, wherein the method comprises the following steps:
(11) the selected palm print images have various noises, and the purpose of denoising the images is to provide a denoised training sample for the subsequent neural network training and perform denoising operation on the palm print images by using a BM3D denoising algorithm;
(12) carrying out filtering enhancement operation on the de-noised palm print image by using Gabor wavelet, wherein Gabor wavelet transform is one of wavelet transform and is orthogonalization of Gabor transform, decomposing the palm print image by using the Gabor wavelet, then carrying out inverse transform, and screening important components by changing corresponding parameters in the process so as to achieve the purpose of image enhancement;
(13) marking the palm print of the training sample, wherein the marked position is a valley point between the middle finger and the other two fingers, and the marked size is x pixels, for example, x is 5, as shown in fig. 1;
(14) marking the palm print of the training sample, wherein the marking position is the region of interest of the palm print, and the marking size is y pixels x z pixels, such as y equals 144, z equals 128, as shown in fig. 1;
(15) performing data expansion on the m palm print images after the operation, wherein the expansion mode is to perform random rotation and left-right turning on the images, and screening the images still retaining the main information of the palm prints to be used as training samples for neural network training;
2) training a multitask convolution neural network by using the training sample generated in the step 1) to obtain a network model for extracting and enhancing the palm print ROI; the method comprises the following steps:
(21) constructing a multitask convolution neural network model to enable the multitask convolution neural network model to simultaneously realize enhancement and segmentation of a palm print image, dividing the network into three parts, wherein the first part is a sharing part, the sharing part carries out down-sampling operation on the palm print image, inputs the image and carries out operation of 2 convolution modules, the size of a convolution kernel is 3 x 3, a characteristic layer is 64, and the output of the last convolution module is stored; performing maximum pooling operation on the result, performing convolution operation of 2 convolution modules, wherein the size of a convolution kernel is 3 x 3, a characteristic layer is 128, and the result is output and stored through the last convolution module; the subsequent processing is the same as above, and the convolution is performed after twice pooling, the convolution kernel of the corresponding convolution module is 3 × 3, the feature layer is 256, the convolution kernel is 3 × 3, and the feature layer is 512, as shown in fig. 2;
(22) the second part is an independent part which is a palm print image segmentation part, the palm print image segmentation part is up-sampled on the basis of the output of the first part, the up-sampling process is 3 times of deconvolution and 3 times of cross operation of 2 convolution modules, and finally a palm print segmentation graph with the same size as the original image is output, the convolution kernel of the convolution layer of the part is 3 x 3, and the sizes of the characteristic layers are 256, 28 and 64 in sequence. The convolution kernel of the last convolution operation is 1 x 1, and the feature level is 1, as shown in fig. 2;
(23) the third part is an independent part and is a palm print image enhancement part. And performing upsampling on the first part of output on the basis of the first part of output, wherein the upsampling process is the cross operation of 3 deconvolution modules and 3 times of 2 convolution modules, and finally outputting a palm print segmentation graph with the same size as the original graph, the convolution kernel of the convolution layer of the first part is 3 x 3, and the sizes of the characteristic layers are 256, 28 and 64 in sequence. The convolution kernel of the last convolution operation is 1 x 1, and the feature level is 1, as shown in fig. 2;
(24) in the constructed multitask convolutional neural network, the activation functions of two shared convolutional layers are both ReLU, in the independent convolutional layers of the two tasks, the palm print ROI extracts the last layer of the network structure to be sigmoid, and the activation function of the palm print image enhancement network structure is ReLU;
(25) in the training process, the total loss function of the multitask convolutional neural network is as follows:
Loss1+2＝α*loss1+β*loss2
loss1+2is the total loss function of the multi-task neural network, alpha and beta are the weight coefficients preset corresponding to each task, loss1、loss2Extracting a loss function and a palm print image loss function for a palm print ROI in the multitask convolution neural network respectively, wherein the process is as follows:
(251) loss function loss for palm print ROI extraction1Is defined as:
(252) loss function loss for palm print image enhancement2Is defined as:
wherein f isiRepresenting a predicted value; y isiRepresenting the true value;
3) verifying the trained multitask convolution neural network model through a verification set, outputting a result, and correcting the result, wherein the method comprises the following steps:
(31) inputting an image to be processed into a trained multitask convolution neural network, comparing an output result of the network with an original image in a test set, calculating a loss function, performing network back propagation until the loss function is stable, and outputting a result graph: a palm print calibration image and a palm print enhancement image;
(32) correcting the calibration graph of the region of interest of the palm print, wherein the correcting steps are as follows:
(321) the two valley points are connected, and a straight line and a corresponding vertical line can be drawn. New coordinate axes will be established with these two straight lines. Taking the intersection point of the two straight lines as an origin, firstly, rotating the palm print image by an angle theta by taking the origin point O as a center, wherein the angle theta is expressed as:
when theta is a positive value, rotating the palm print image anticlockwise; otherwise, clockwise rotating the palm print image, and judging the distance between the palm print interested region automatically calibrated by the neural network and the valley point by taking the two valley points as reference objects;
(322) according to experience judgment, the distance between the left side boundary of the palm print interesting region and the valley point connecting line can be determined to be one fourth of the distance between the two valley points, the distance between the right side boundary is 1.5 times of the distance between the two valley points, if the condition is met, the calibration result does not need to be corrected, otherwise, correction processing needs to be carried out;
(323) if the palm print calibration result is not satisfied, automatically correcting the palm print calibration result, filling a small part of the palm print interesting region, and restoring the redundant part.
Claims (4)
1. A palm print ROI extraction and enhancement method based on a multitask convolutional neural network is characterized by comprising the following steps:
1) preparing a sample before training, namely firstly copying the sample into two parts, namely A and B; performing BM3D denoising and Gabor wavelet filtering on the sample A so as to perform image enhancement processing on the sample; marking the sample B, and respectively marking two interphalangeal valley points and a palm print ROI area of the training sample; finally, performing data expansion on the marked training sample;
2) training a multitask convolution neural network by using the training sample generated in the step 1) to obtain a network model for extracting and enhancing the palm print ROI;
3) and verifying the trained multi-task convolutional neural network model through the verification set, outputting a result and correcting the result.
2. The method for palm print ROI extraction and enhancement based on multitask convolutional neural network according to claim 1, wherein said step 1) comprises the following steps:
(11) the selected palm print images have various noises, and the purpose of denoising the images is to provide a denoised training sample for the subsequent neural network training and perform denoising operation on the palm print images by using a BM3D denoising algorithm;
(12) carrying out filtering enhancement operation on the de-noised palm print image by using Gabor wavelet, wherein Gabor wavelet transform is one of wavelet transform and is orthogonalization of Gabor transform, decomposing the palm print image by using the Gabor wavelet, then carrying out inverse transform, and screening important components by changing corresponding parameters in the process so as to achieve the purpose of image enhancement;
(13) marking the palm print of the training sample, wherein the marked position is a valley point between the middle finger and the other two fingers, and the marked size is x pixels;
(14) marking the palm print of the training sample, wherein the marking position is a palm print interesting region, and the marking size is y pixels x z pixels, such as y equals 144, z equals 128;
(15) and performing data expansion on the m palm print images after the operation, wherein the expansion mode is to perform random rotation and left-right turning on the images, and screen the images still retaining the main information of the palm prints to be used as training samples for neural network training.
3. The method for palm print ROI extraction and enhancement based on multitask convolutional neural network according to claim 1 or 2, wherein the step 2) comprises the following steps:
(21) constructing a multitask convolution neural network model to enable the multitask convolution neural network model to simultaneously realize enhancement and segmentation of a palm print image, dividing the network into three parts, wherein the first part is a sharing part, the sharing part carries out down-sampling operation on the palm print image, inputs the image and carries out operation of 2 convolution modules, the size of a convolution kernel is 3 x 3, a characteristic layer is 64, and the output of the last convolution module is stored; performing maximum pooling operation on the result, performing convolution operation of 2 convolution modules, wherein the size of a convolution kernel is 3 x 3, a characteristic layer is 128, and the result is output and stored through the last convolution module; the subsequent processing is the same as the above, and the convolution is performed after twice pooling, the convolution kernel of the corresponding convolution module is 3 x 3, the characteristic layer is 256, the convolution kernel is 3 x 3, and the characteristic layer is 512;
(22) the second part is an independent part and is a palm print image segmentation part, the palm print image segmentation part is subjected to upsampling on the basis of the output of the first part, the upsampling process is 3 times of deconvolution and 3 times of cross operation of 2 convolution modules, and finally a palm print segmentation graph with the same size as the original image is output, the convolution kernel of the convolution layer of the part is 3 x 3, the sizes of the feature layers are 256, 28 and 64 in sequence, the convolution kernel of the last convolution operation is 1 x 1, and the feature layer is 1;
(23) the third part is an independent part and is a palm print image enhancement part, the palm print image enhancement part is subjected to upsampling on the basis of the output of the first part, the upsampling process is the cross operation of 3 deconvolution modules and 3 convolution modules, and finally a palm print segmentation graph with the same size as the original image is output, the convolution kernel of the convolution layer of the part is 3 x 3, the sizes of the characteristic layers are 256, 28 and 64 in sequence, the convolution kernel of the last convolution operation is 1 x 1, and the characteristic layer is 1;
(24) in the constructed multitask convolutional neural network, the activation functions of two shared convolutional layers are both ReLU, in the independent convolutional layers of the two tasks, the palm print ROI extracts the last layer of the network structure to be sigmoid, and the activation function of the palm print image enhancement network structure is ReLU;
(25) in the training process, the total loss function of the multitask convolutional neural network is as follows:
Loss1+2＝α*loss1+β*loss2
Loss1+2is the total loss function of the multi-task neural network, alpha and beta are the weight coefficients preset corresponding to each task, loss1、loss2Extracting a loss function and a palm print image loss function for a palm print ROI in the multitask convolution neural network respectively, wherein the process is as follows:
(251) loss function loss for palm print ROI extraction1Is defined as:
(252) loss function loss for palm print image enhancement2Is defined as:
wherein f isiRepresenting a predicted value; y isiRepresenting the true value.
4. The method for palm print ROI extraction and enhancement based on multitask convolutional neural network according to claim 1 or 2, characterized in that said step 3) comprises the following steps:
(31) inputting an image to be processed into a trained multitask convolution neural network, comparing an output result of the network with an original image in a test set, calculating a loss function, performing network back propagation until the loss function is stable, and outputting a result graph: a palm print calibration image and a palm print enhancement image;
(32) correcting the calibration graph of the region of interest of the palm print, wherein the correcting steps are as follows:
(321) connecting two valley points, drawing a straight line and a corresponding vertical line, establishing a new coordinate axis by using the two straight lines, taking the intersection point of the two straight lines as an original point, rotating the palm print image by taking the original point O as a center by an angle theta, and expressing the angle theta as follows:
when theta is a positive value, rotating the palm print image anticlockwise; otherwise, clockwise rotating the palm print image, and judging the distance between the palm print interested region automatically calibrated by the neural network and the valley point by taking the two valley points as reference objects;
(322) according to experience judgment, the distance between the left side boundary of the palm print interesting region and the valley point connecting line can be determined to be one fourth of the distance between the two valley points, the distance between the right side boundary is 1.5 times of the distance between the two valley points, if the condition is met, the calibration result does not need to be corrected, otherwise, correction processing needs to be carried out;
(323) if the palm print calibration result is not satisfied, automatically correcting the palm print calibration result, filling a small part of the palm print interesting region, and restoring the redundant part.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010916060.3A CN112149521B (en) | 2020-09-03 | 2020-09-03 | Palm print ROI extraction and enhancement method based on multitasking convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010916060.3A CN112149521B (en) | 2020-09-03 | 2020-09-03 | Palm print ROI extraction and enhancement method based on multitasking convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112149521A true CN112149521A (en) | 2020-12-29 |
| CN112149521B CN112149521B (en) | 2024-05-07 |
Family
ID=73889268
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010916060.3A Active CN112149521B (en) | 2020-09-03 | 2020-09-03 | Palm print ROI extraction and enhancement method based on multitasking convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112149521B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191987A (en) * | 2021-05-31 | 2021-07-30 | 齐鲁工业大学 | Palm print image enhancement method based on PCNN and Otsu |
| CN114140424A (en) * | 2021-11-29 | 2022-03-04 | 佳都科技集团股份有限公司 | Palm vein data enhancement method and device, electronic equipment and medium |
| WO2023160048A1 (en) * | 2022-02-28 | 2023-08-31 | 腾讯科技（深圳）有限公司 | Palmprint sample generation method and apparatus, and device, medium and program product |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105701513A (en) * | 2016-01-14 | 2016-06-22 | 深圳市未来媒体技术研究院 | Method of rapidly extracting area of interest of palm print |
| CN107977671A (en) * | 2017-10-27 | 2018-05-01 | 浙江工业大学 | A kind of tongue picture sorting technique based on multitask convolutional neural networks |
| CN108564097A (en) * | 2017-12-05 | 2018-09-21 | 华南理工大学 | A kind of multiscale target detection method based on depth convolutional neural networks |
| KR20200046169A (en) * | 2018-10-17 | 2020-05-07 | 엔에이치엔 주식회사 | Neural network system for detecting Palmprint and method for providing Palmprint-based fortune forecasting service |
-
2020
- 2020-09-03 CN CN202010916060.3A patent/CN112149521B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105701513A (en) * | 2016-01-14 | 2016-06-22 | 深圳市未来媒体技术研究院 | Method of rapidly extracting area of interest of palm print |
| CN107977671A (en) * | 2017-10-27 | 2018-05-01 | 浙江工业大学 | A kind of tongue picture sorting technique based on multitask convolutional neural networks |
| CN108564097A (en) * | 2017-12-05 | 2018-09-21 | 华南理工大学 | A kind of multiscale target detection method based on depth convolutional neural networks |
| KR20200046169A (en) * | 2018-10-17 | 2020-05-07 | 엔에이치엔 주식회사 | Neural network system for detecting Palmprint and method for providing Palmprint-based fortune forecasting service |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113191987A (en) * | 2021-05-31 | 2021-07-30 | 齐鲁工业大学 | Palm print image enhancement method based on PCNN and Otsu |
| CN114140424A (en) * | 2021-11-29 | 2022-03-04 | 佳都科技集团股份有限公司 | Palm vein data enhancement method and device, electronic equipment and medium |
| CN114140424B (en) * | 2021-11-29 | 2023-07-18 | 佳都科技集团股份有限公司 | Palm vein data enhancement method, palm vein data enhancement device, electronic equipment and medium |
| WO2023160048A1 (en) * | 2022-02-28 | 2023-08-31 | 腾讯科技（深圳）有限公司 | Palmprint sample generation method and apparatus, and device, medium and program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112149521B (en) | 2024-05-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409222B (en) | Multi-view facial expression recognition method based on mobile terminal | |
| CN109685819B (en) | Three-dimensional medical image segmentation method based on feature enhancement | |
| CN106529447B (en) | Method for identifying face of thumbnail | |
| CN112149521A (en) | Palm print ROI extraction and enhancement method based on multitask convolutional neural network | |
| CN110059593B (en) | Facial expression recognition method based on feedback convolutional neural network | |
| CN115457568B (en) | Historical document image noise reduction method and system based on generation countermeasure network | |
| CN110245621B (en) | Face recognition device, image processing method, feature extraction model, and storage medium | |
| CN110674824A (en) | Finger vein segmentation method and device based on R2U-Net and storage medium | |
| CN113658040A (en) | Face super-resolution method based on prior information and attention fusion mechanism | |
| CN111080591A (en) | Medical image segmentation method based on combination of coding and decoding structure and residual error module | |
| CN109872326B (en) | Contour detection method based on deep reinforced network jump connection | |
| CN109376859A (en) | A kind of neural networks pruning method based on diamond shape convolution | |
| CN113362310A (en) | Medical image liver segmentation method based on unsupervised learning | |
| CN106709431A (en) | Iris recognition method and device | |
| CN116051408B (en) | Image depth denoising method based on residual error self-coding | |
| CN113554084A (en) | Vehicle re-identification model compression method and system based on pruning and light-weight convolution | |
| EP3671635B1 (en) | Curvilinear object segmentation with noise priors | |
| CN112785539A (en) | Multi-focus image fusion method based on image adaptive decomposition and parameter adaptive | |
| Guo et al. | Multifeature extracting CNN with concatenation for image denoising | |
| CN109949334B (en) | Contour detection method based on deep reinforced network residual error connection | |
| CN113221660B (en) | Cross-age face recognition method based on feature fusion | |
| Chen et al. | A finger vein recognition algorithm based on deep learning | |
| CN116934613A (en) | Branch convolution channel attention module for character repair | |
| CN112990123B (en) | Image processing method, apparatus, computer device and medium | |
| CN112801909B (en) | Image fusion denoising method and system based on U-Net and pyramid module |
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 | ||
| GR01 | Patent grant |