# US 2021/0174104 A1 · Finger vein comparison method, computer equipment, and storage medium

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0174104 A1
- Document ID: US20210174104A1
- Status: Published application
- Kind: A1
- Application number: 17/178,911
- Filed: 2021-02-18
- Published: 2021-06-10
- Earliest claimed: 2019-04-03
- Assignees: Ping An Technology (Shenzhen) Co., Ltd.
- Applicants: Ping An Technology (Shenzhen) Co., Ltd.
- Inventors: Zhongdi CHAO; Bojin ZHUANG; Shaojun WANG
- Web page: https://patentgrep.com/patent/US20210174104A1

## CPC

- G06V 10/764: using classification, e.g. of video objects
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06V 10/82: using neural networks
- G06V 40/10: Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V 40/14: Vascular patterns
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A finger vein comparison method, a computer equipment, and a storage medium are provided. The finger vein comparison method includes: two finger vein images to be compared are obtained (S10); image channel fusion is performed on the two finger vein images to be compared to obtain a two-channel target finger vein image to be compared (S20); the target finger vein image to be compared is input into a feature extractor, and a feature vector of the target finger vein image to be compared is extracted by the feature extractor (S30); the feature vector of the target finger vein image to be compared is input into a dichotomy classifier to obtain a dichotomy result (S40); and it is determined according to the dichotomy result whether the two finger vein images to be compared come from the same finger (S50).

## Claims

### Claim 1

A finger vein comparison method, comprising:
obtaining two finger vein images to be compared, wherein the two finger vein images to be compared are single-channel images;
performing image channel fusion on the two finger vein images to be compared to obtain a two-channel target finger vein image to be compared;
inputting the target finger vein image to be compared into a feature extractor, and extracting a feature vector of the target finger vein image to be compared by the feature extractor, wherein the feature extractor is trained by using a two-channel convolutional neural network;
inputting the feature vector of the target finger vein image to be compared into a dichotomy classifier to obtain a dichotomy result; and
determining according to the dichotomy result whether the two finger vein images to be compared come from a same finger.

### Claim 2 (depends on 1)

The method as claimed in claim 1, wherein performing image channel fusion on the two finger vein images to be compared to obtain the two-channel target finger vein image to be compared comprises:
obtaining a single-channel pixel matrix of a first finger vein image to be compared;
obtaining a single-channel pixel matrix of a second finger vein image to be compared; and
performing image channel fusion on pixel values of corresponding positions in the single-channel pixel matrix of the first finger vein image to be compared and in the single-channel pixel matrix of the second finger vein image to be compared, to obtain the target finger vein image to be compared.

### Claim 3 (depends on 1)

The method as claimed in claim 1, further comprising before inputting the target finger vein image to be compared into the feature extractor:
obtaining a finger vein sample to be trained;
inputting the finger vein sample to be trained into the two-channel convolutional neural network for a training process, and using an enhanced edge cosine loss function to calculate a loss value generated in the training process, wherein the two-channel convolutional neural network comprises four convolution layers, four pooling layers, and one fully connected layer, wherein the first layer, the third layer, the fifth layer, and the seventh layer of the two-channel convolutional neural network are the convolution layers, wherein the second layer, the fourth layer, the sixth layer, and the eighth layer are the pooling layers, and wherein the ninth layer is the fully connected layer; and
updating the two-channel convolutional neural network according to the loss value to obtain the feature extractor.

### Claim 4 (depends on 3)

The method as claimed in claim 3, wherein obtaining the finger vein sample to be trained comprises:
obtaining single-channel finger vein images to be trained;
performing image channel fusion on every two single-channel finger vein images to be trained to obtain a two-channel finger vein image to be trained;
when the single-channel finger vein images to be trained for fusing come from the same finger, setting a label value of the two-channel finger vein image to be trained to 0;
when the single-channel finger vein images to be trained for fusing do not come from the same finger, setting the label value of the two-channel finger vein image to be trained to 1; and
obtaining the finger vein sample to be trained according to the set label value.

### Claim 5 (depends on 1)

The method as claimed in claim 1, wherein determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger comprises:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 6 (depends on 2)

The method as claimed in claim 2, wherein determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger comprises:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 7 (depends on 3)

The method as claimed in claim 3, wherein determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger comprises:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 8

Computer equipment, comprising:
a memory, a processor, and a computer readable instruction stored in the memory and capable of running on the processor, wherein the processor, when executing the computer readable instruction, implements:
obtaining two finger vein images to be compared, wherein the two finger vein images are single-channel images;
performing image channel fusion on the two finger vein images to be compared to obtain a two-channel target finger vein image to be compared;
inputting the target finger vein image to be compared into a feature extractor, and extracting a feature vector of the target finger vein image to be compared by the feature extractor, wherein the feature extractor is trained by using a two-channel convolutional neural network;
inputting the feature vector of the target finger vein image to be compared into a dichotomy classifier to obtain a dichotomy result; and
determining according to the dichotomy result whether the two finger vein images to be compared come from a same finger.

### Claim 9 (depends on 8)

The computer equipment as claimed in claim 8, wherein to implement performing image channel fusion on the two finger vein images to be compared to obtain the two-channel target finger vein image to be compared, the processor executes the computer readable instruction to implement:
obtaining a single-channel pixel matrix of a first finger vein image to be compared;
obtaining a single-channel pixel matrix of a second finger vein image to be compared; and
performing image channel fusion on pixel values of corresponding positions in the single-channel pixel matrix of the first finger vein image to be compared and in the single-channel pixel matrix of the second finger vein image to be compared, to obtain the target finger vein image to be compared.

### Claim 10 (depends on 8)

The computer equipment as claimed in claim 8, wherein the processor, when executing the computer readable instruction, further implements:
before inputting the target finger vein image to be compared into the feature extractor, obtaining a finger vein sample to be trained;
inputting the finger vein sample to be trained into the two-channel convolutional neural network for a training process, and using an enhanced edge cosine loss function to calculate a loss value generated in the training process, wherein the two-channel convolutional neural network comprises four convolution layers, four pooling layers, and one fully connected layer, wherein the first layer, the third layer, the fifth layer, and the seventh layer of the two-channel convolutional neural network are the convolution layers, wherein the second layer, the fourth layer, the sixth layer, and the eighth layer are the pooling layers, and wherein the ninth layer is the fully connected layer; and
updating the two-channel convolutional neural network according to the loss value to obtain the feature extractor.

### Claim 11 (depends on 10)

The computer equipment as claimed in claim 10, wherein to implement obtaining the finger vein sample to be trained, the processor executes the computer readable instruction to implement:
obtaining single-channel finger vein images to be trained;
performing image channel fusion on every two single-channel finger vein images to be trained to obtain a two-channel finger vein image to be trained;
when the single-channel finger vein images to be trained for fusing come from the same finger, setting a label value of the two-channel finger vein image to be trained to 0;
when the single-channel finger vein images to be trained for fusing do not come from the same finger, setting the label value of the two-channel finger vein image to be trained to 1; and
obtaining the finger vein sample to be trained according to the set label value.

### Claim 12 (depends on 8)

The computer equipment as claimed in claim 8, wherein to implement determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger, the processor executes the computer readable instruction to implement:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 13 (depends on 9)

The computer equipment as claimed in claim 9, wherein to implement determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger, the processor executes the computer readable instruction to implement:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 14 (depends on 10)

The computer equipment as claimed in claim 10, wherein to implement determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger, the processor executes the computer readable instruction to implement:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 15

A non-transitory computer readable storage medium that stores a computer readable instruction, wherein the computer readable instruction, when executed by a processor, enables the processor to implement:
obtaining two finger vein images to be compared are obtained, wherein the two finger vein images are single-channel images;
performing image channel fusion on the two finger vein images to be compared to obtain a two-channel target finger vein image to be compared;
inputting the target finger vein image to be compared into a feature extractor, and extracting a feature vector of the target finger vein image to be compared by the feature extractor, wherein the feature extractor is trained by using a two-channel convolutional neural network;
inputting the feature vector of the target finger vein image to be compared into a dichotomy classifier to obtain a dichotomy result; and
determining according to the dichotomy result whether the two finger vein images to be compared come from a same finger.

### Claim 16 (depends on 15)

The non-transitory computer readable storage medium as claimed in claim 15, to implement performing image channel fusion on the two finger vein images to be compared to obtain the two-channel target finger vein image to be compared, the computer readable instruction is executed by the processor to implement:
obtaining a single-channel pixel matrix of a first finger vein image to be compared;
obtaining a single-channel pixel matrix of a second finger vein image to be compared; and
performing image channel fusion on pixel values of corresponding positions in the single-channel pixel matrix of the first finger vein image to be compared and in the single-channel pixel matrix of the second finger vein image to be compared, to obtain the target finger vein image to be compared.

### Claim 17 (depends on 15)

The non-transitory computer readable storage medium as claimed in claim 15, wherein the computer readable instruction, when executed by the processor, enables the processor to further implement:
before inputting the target finger vein image to be compared into the feature extractor, obtaining a finger vein sample to be trained;
inputting the finger vein sample to be trained into the two-channel convolutional neural network for a training process, and using an enhanced edge cosine loss function to calculate a loss value generated in the training process, wherein the two-channel convolutional neural network comprises four convolution layers, four pooling layers, and one fully connected layer, wherein the first layer, the third layer, the fifth layer, and the seventh layer of the two-channel convolutional neural network are the convolution layers, wherein the second layer, the fourth layer, the sixth layer, and the eighth layer are the pooling layers, and wherein the ninth layer is the fully connected layer; and
updating the two-channel convolutional neural network according to the loss value to obtain the feature extractor.

### Claim 18 (depends on 17)

The non-transitory computer readable storage medium as claimed in claim 17, wherein to implement obtaining the finger vein sample to be trained, the computer readable instruction is executed by the processor to implement:
obtaining single-channel finger vein images to be trained;
performing image channel fusion on every two single-channel finger vein images to be trained to obtain a two-channel finger vein image to be trained;
when the single-channel finger vein images to be trained for fusing come from the same finger, setting a label value of the two-channel finger vein image to be trained to 0;
when the single-channel finger vein images to be trained for fusing do not come from the same finger, setting the label value of the two-channel finger vein image to be trained to 1; and
obtaining the finger vein sample to be trained according to the set label value.

### Claim 19 (depends on 15)

The non-transitory computer readable storage medium as claimed in claim 15, wherein to implement determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger, the computer readable instruction is executed by the processor to implement:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.

### Claim 20 (depends on 16)

The non-transitory computer readable storage medium as claimed in claim 16, wherein to implement determining according to the dichotomy result whether the two finger vein images to be compared come from the same finger, the computer readable instruction is executed by the processor to implement:
when a value of the dichotomy result is 0, determining that the two finger vein images to be compared come from the same finger; and
when the value of the dichotomy result is 1, determining that the two finger vein images to be compared do not come from the same finger.
