CN112022093A - A skin imaging system - Google Patents

Abstract

The invention discloses a skin imaging system, which comprises an OCT subsystem, an HSI subsystem and an image processing unit, wherein the OCT subsystem is used for emitting a light signal, the light signal is reflected to form reference light and sample light of a skin area, the sample light and the reference light are interfered to form interference spectrum information, the HSI subsystem is used for emitting a white light signal to the skin area, reflected light generated by the white light signal irradiating the skin area is split and converged to obtain a spectrogram in all wavelength ranges, the image processing unit is used for generating a depth image of the skin area according to the interference spectrum information, reconstructing and correcting the spectrogram to obtain a hyperspectral image of the skin area, and the depth image and the hyperspectral image are fused to obtain the skin image. The invention utilizes the double imaging of the OCT technology and the hyperspectral technology, not only can make up the defects of the respective imaging technologies, but also can realize advantage complementation.

Description

A skin imaging system

【Technical field】

The invention relates to the field of biomedical imaging, in particular to a skin imaging system.

【Background technique】

Optical coherence tomography (Optical Coherence Tomography, OCT) and hyperspectral imaging (Hyper-spectral Imaging, HSI) are two new biomedical optical imaging technologies, which can achieve non-destructive imaging of biological tissues in vivo, with high sensitivity, High resolution, non-invasive and radiation-free. These two technologies have been widely used in the field of biomedicine and have become important means of clinical outpatient examination. However, in the actual diagnosis process, the two techniques have imaging defects, such as the contradiction between image resolution and imaging range and the inability to obtain molecular information in OCT imaging. The defect of poor ability to obtain depth information.

To this end, we propose a skin imaging system.

[Content of the invention]

The main purpose of the present invention is to provide a skin imaging system. The OCT system used for biomedical imaging and the HSI system are combined to form a dual-modal imaging system, which realizes the complementary advantages of the OCT system and the HSI system, and effectively solves the above-mentioned technology. problems in the background.

To achieve the above object, the technical scheme adopted in the present invention is:

A skin imaging system includes an OCT subsystem, an HSI subsystem and an image processing unit;

OCT subsystem: used to emit light signals, and make the light signals form reference light and sample light in the skin area through reflection, and at the same time cause the sample light and reference light to interfere to form interference spectral information;

The HSI subsystem is used to emit white light signals to the skin area, and to split and condense the reflected light generated by the white light signals irradiating the skin area to obtain spectral maps in all wavelength ranges;

Image processing unit: used to generate a depth image of the skin area according to the interference spectral information, reconstruct and correct the spectral map to obtain a hyperspectral image of the skin area, and fuse the depth image and the hyperspectral image to obtain a skin image.

Optionally, the OCT subsystem includes:

Sweep frequency light source: used to emit laser signals;

The first fiber coupler: used to split the laser signal output by the frequency sweep light source, and make the split optical signal enter the reference arm and the sample arm respectively;

Reference arm: used to reflect the received incident light to generate reference light;

Sample arm: used to emit the received incident light to the skin area to be imaged and form sample light after reflection;

The second fiber coupler: used to interfere the sample light and the reference light to form interference spectral information;

Balanced detector: used to detect interference spectrum information and output to the image processing unit.

Optionally, the sample arm includes a scanning galvanometer and a first achromatic lens, and the incident light entering the sample arm is commutated by the scanning galvanometer and then focused on the skin area through the first achromatic lens.

Optionally, the scanning galvanometer is equipped with a high-speed analog output device, and the high-speed analog output device can output an analog signal to control the scanning galvanometer to scan the skin area.

Optionally, the front end of the scanning galvanometer is further configured with a first optical fiber collimator, which is used for collimating the optical signal incident on the scanning galvanometer.

Optionally, the reference arm includes a second fiber collimator, a third fiber collimator, a first plane mirror and a second plane mirror, and the incident light entering the reference arm is collimated by the second fiber collimator and then passed through the first plane mirror. , The second plane mirror enters the third fiber collimator after reflection.

Optionally, the HSI subsystem includes:

White light source: used to emit white light signals to the skin area;

Front optical system: It is used to achromatize the reflected light generated by the white light signal irradiating the skin area, and through focusing processing, the reflected light can pass through the preset incident slit;

Spectrometer: used for push-brooming according to the preset step size, and performing spectral splitting and convergence processing on the reflected light passing through the incident slit to obtain the spectrum in all wavelength ranges of the skin area; where the step size is equal to the width of the incident slit;

CCD camera: used to detect the spectrum output by the spectrometer to form a spectrogram.

Optionally, the spectrometer is configured with a motorized translation stage for driving the spectrometer to run to complete the push-broom operation.

Optionally, the front optical system includes at least two second achromatic lenses arranged coaxially.

Optionally, a dichroic mirror is also included, and the dichroic mirror is used to focus the optical signals emitted by the OCT subsystem and the HSI subsystem on the skin area, and at the same time, it is used to reflect the optical signals reflected from the skin area to the OCT subsystem. and HSI subsystem.

Compared with the prior art, the present invention has the following beneficial effects:

This skin imaging system includes OCT subsystem, HSI subsystem and image processing unit. Using the dual imaging of OCT technology and hyperspectral technology, it can not only make up for the defects of the respective imaging technologies, but also achieve complementary advantages. , and can obtain a larger imaging depth. In the actual diagnosis process, the skin imaging system makes up for the imaging defects of the OCT subsystem and the HSI subsystem respectively, which is beneficial to accurately determine the surface boundary of the lesion area. And analyze the difference between diseased tissue and normal tissue through deep structural information.

【Description of drawings】

The accompanying drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention.

FIG. 1 is an overall structural diagram of an embodiment of a skin imaging system.

FIG. 2 is a pigmented nevus sample in the embodiment of the skin imaging system.

FIG. 3 is a comparison diagram before and after spectral denoising of normal skin in an embodiment of the skin imaging system.

FIG. 4 is an actual measurement diagram of the spectral resolution of the HSI subsystem in the embodiment of the skin imaging system.

FIG. 5 is a result and comparison diagram of various clustering algorithms in an embodiment of the skin imaging system.

FIG. 6 is a 3D reconstruction diagram of a pigmented nevus sample in an embodiment of the skin imaging system.

FIG. 7 is a B-scan diagram of the skin in the embodiment of the skin imaging system and a B-scan diagram after flattening.

FIG. 8 is an A-scan comparison diagram of normal skin and a pigmented nevus sample in an embodiment of the skin imaging system.

FIG. 9 is a fitting result of the attenuation coefficient in the embodiment of the skin imaging system.

FIG. 10 is a B-scan diagram of the OCT subsystem in the embodiment of the skin imaging system.

In the figure: 1. Frequency sweep light source; 2. The first fiber coupler; 3. The second fiber coupler; 4. The first fiber collimator; 5, The first plane mirror; 6, The second plane mirror; 7, The second Fiber collimator; 8. Balance detector; 9. Third fiber collimator; 10. Scanning mirror; 11. First achromatic lens; 12. Sample; 13. White light source; 14. Second achromatic lens 15. Front optical system; 16. Incident slit; 17. Spectrometer; 18. CCD camera; 19. Image processing unit; 20. Dichroic mirror.

【Detailed ways】

In order to make the technical means, creative features, goals and effects realized by the present invention easy to understand, the technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only some, but not all, embodiments of the present invention. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

The following OCT subsystem refers to the optical coherence tomography imaging subsystem, and the HSI subsystem refers to the hyperspectral imaging subsystem.

As shown in FIG. 1 , the skin imaging system provided by the present invention includes an OCT subsystem, an HSI subsystem and an image processing unit 19;

The OCT subsystem emits an optical signal, and the optical signal is reflected to form the reference light and the sample light of the skin area, and at the same time, the sample light and the reference light interfere to form the interference spectrum information;

The HSI subsystem emits a white light signal to the skin area, and divides and condenses the reflected light generated by the white light signal irradiating the skin area to obtain spectrograms in all wavelength ranges;

On the one hand, the image processing unit 19 generates a depth image of the skin area according to the interference spectral information formed by the OCT subsystem, and on the other hand, performs reconstruction and correction according to the spectral map obtained by the HSI subsystem to obtain a hyperspectral image of the skin area, and finally, converts the depth image into the skin area. Fusion with hyperspectral images to obtain skin images.

The operation of the OCT subsystem and the HSI subsystem is not limited in sequence. It can be that the OCT subsystem runs first and the HSI subsystem runs later; or the HSI subsystem runs first and the OCT subsystem runs after, or the OCT subsystem and the HSI subsystem run. run simultaneously.

In some embodiments, the OCT subsystem includes a swept frequency light source 1, a reference arm, a sample arm, a first fiber coupler 2, a second fiber coupler 3, and a balanced detector 8;

In this embodiment, the model of the swept frequency light source 1 can be SS-OCT1060, Axsun Technology Inc, the center wavelength of the light source is 1060nm, the bandwidth is 110nm, and the frequency sweep rate is 100kHz; the model of the balanced detector 8 can be PDB471C, Thorlabs; It should be noted that the models of the frequency sweep light source 1 and the balanced detector 8 are not limited to the models of this embodiment.

First, the light emitted by the swept frequency light source 1 is split through the first fiber coupler 2; after that, it enters the sample arm (50%) and the reference arm (50%) respectively, and the incident light received by the reference arm is reflected to form a reference light , after the incident light received by the sample arm changes direction, the skin area to be imaged is focused and reflected to form sample light; finally, the sample light in the sample arm and the reference light in the reference arm are in the second fiber coupler 3 Optical interference occurs to form interference spectrum information, the balance detector 8 detects the interference spectrum information and outputs the interference spectrum information to the image processing unit 19 for processing. Specifically, the interference spectrum signal detected in the balance detector 8 first passes through a A low-pass filter (model can be SLP-150+, Mini-circuits) to filter out other signals with frequencies higher than 155MHz, and then output to the image processing unit through a 12-bit dual-channel data acquisition card (model ATS9351, Alazartech) 19 .

In some embodiments, the sample arm includes a scanning galvanometer 10 and a first achromatic lens 11;

The model of the scanning galvanometer 10 can be GVSM002/M, Thorlabs; the model of the first achromatic lens 11 can be AC254-125-B-ML, Thorlabs;

The light entering the sample arm changes direction through the scanning galvanometer 10 , and then passes through the first achromatic lens 11 to achieve focusing on the skin area.

In some embodiments, the scanning galvanometer 10 is configured with a high-speed analog output device;

The model of the high-speed analog output device can be PCIe-6363, National Instruments;

The high-speed analog output device can output analog signals to control and drive the scanning galvanometer 10 to realize scanning in the X and Y directions, and finally analyze and visualize the data through the image processing unit 19 .

In some embodiments, the front end of the scanning galvanometer 10 configured on the sample arm is further configured with a first optical fiber collimator 4 , and the first optical fiber collimator 4 is used for collimating the optical signal incident on the scanning galvanometer 10 .

In some embodiments, the reference arm includes a second fiber collimator 7 , a third fiber collimator 9 , a first plane mirror 5 and a second plane mirror 6 , and the incident light entering the reference arm passes through the second fiber collimator 7 After collimation, it enters the third fiber collimator 9 after being reflected by the first plane mirror 5 and the second plane mirror 6 .

The above-mentioned incident light incident on the reference arm passes through the second fiber collimator 7 , the first plane mirror 5 , the second plane mirror 6 , and the third fiber collimator 9 in sequence to change the direction of the incident light.

In some embodiments, the HSI subsystem includes a white light source 13, a front optical system 15, a spectrometer 17, and a CCD camera 18;

The white light source 13 emits a white light signal to the skin area, and the reflected light generated by the front optical system 15 irradiating the white light signal to the skin area is achromatic, and the reflected light can pass through the preset incident slit 16 of the spectrometer 17 through focusing processing;

The spectrometer 17 is installed on the mobile device and push-broom is performed based on a preset step size. At the same time, the spectrometer 17 performs spectral splitting and convergence processing on the reflected light passing through the incident slit 16 to obtain the spectrum in all wavelength ranges of the skin area;

The above-mentioned step size is equal to the width of the incident slit 16, so that each step of the push-broom of the spectrometer 17 will obtain the spectrogram of the skin region in all wavelength ranges;

The CCD detection camera collects the above-mentioned spectrograms, and through the later data processing on the computer, the biological tissue images at all wavelengths are obtained after reconstruction and correction;

In this embodiment, two 60W wide-beam LED spotlights can be used as the white light source 13, and the model of the CCD detection camera is MANTA G-201-30FPS, Allied Vision.

In some embodiments, the spectrometer 17 is configured with a motorized translation stage for driving the spectrometer 17 to run to complete the push-broom operation, specifically, the spectrometer 17 is fixed on the motorized translation stage to push-broom it.

In some embodiments, the front optical system 15 includes at least two second achromatic lenses 14 arranged on a coaxial line. The second achromatic lenses 14 achromat and focus the reflected light generated by the white light signal irradiating the skin area, so that the The reflected light can pass through the predetermined incident slit 16 .

In this embodiment, the width of the above-mentioned slit is 22 μm.

In some embodiments, the skin imaging system further includes a dichroic mirror 20, the dichroic mirror 20 focuses the optical signals emitted by the OCT subsystem and the HSI subsystem on the skin area, and simultaneously reflects the optical signals reflected from the skin area to OCT subsystem and HSI subsystem;

The imaging performance parameters of this system are as follows: the A-scan speed of the OCT system reaches 100 kHz, the imaging depth is 5 mm (in air), and the axial and lateral resolutions are 8 μm and 70.3 μm, respectively; the imaging spectral range of the HSI system is 465-630 nm, and the spectral resolution The rate and spatial resolution are 2.1 nm and 31.3 μm, respectively. For details, see Figure 4, the measured map of the spectral resolution of the hyperspectral subsystem.

The image processing unit 19 includes basic information processing equipment such as a computer;

In this system, the image processing unit 19 includes timing control, data acquisition, image reconstruction, real-time image display and image preprocessing of the OCT subsystem and the HSI subsystem, and the software of the image processing unit 19 is partially based on the mixed programming of Labview and Matlab.

The parameters of data acquisition are set on the front panel of Labview, and the B-scan image and original interference spectrum of the OCT part and the spectral image of the hyperspectral part are displayed in real time at the same time. Image preprocessing mainly includes spectral shaping of the OCT subsystem, Fourier Leaf transformation, fixed pattern noise removal, etc. and wavelength calibration of hyperspectral subsystems.

Specific experiments and their analysis:

This experiment is for the detection and research analysis of skin pigmented moles;

The sample of the experiment is shown in Figure 2. The size of the hyperspectral image is 240 × 150 pixels, the spectral dimension is 900 dimensions, and the field of view is 5.28 mm × 3.3 mm. The size of the OCT En-face image is 2000 × 300 pixels. The field size is 12mm x 4mm.

Perform smoothing denoising, multivariate scattering correction, dimensionality reduction, and cluster analysis on hyperspectral data;

The smoothing denoising method adopts Gaussian weighted filter, the smoothing window is set to 5, and the result is shown in Figure 3; Show);

The clustering methods are K-means, GMM and HAC algorithms. The processing results obtained are shown in Figure 5 and Table 2. The system data can obtain more than 90% accuracy and high Dice coefficient under the three algorithms. Among them, the accuracy rate and Dice coefficient obtained by HAC algorithm are the highest, the accuracy rate is 92.61%, and the Dice coefficient is 94.53%.

In Figure 5:

Among them, (a) gold standard; (b) result; (c) comparison;

Among them, 1 represents the K-means algorithm; 2 represents the GMM algorithm; 3 represents the HAC algorithm.

Table 1

main ingredient Contribution rate Cumulative contribution rate 1 0.5679 0.5679 2 0.1442 0.7121 3 0.0638 0.7759 … … … 99 0.006 0.8995 100 0.006 0.9001

Table 2

Accuracy Sensitivity specificity false negative rate false positive rate Dice coefficient K-means 0.9023 0.8529 0.9354 0.1471 0.0646 0.8751 GMM 0.9182 0.8793 0.9443 0.1207 0.0557 0.9045 HAC 0.9261 0.9373 0.8852 0.0627 0.1148 0.9453

On the other hand, a three-dimensional image of the detection area was reconstructed using OCT data (Fig. 6) and the difference between the pigmented nevus area and the normal skin area was quantitatively analyzed by attenuation coefficient and fractal dimension.

In Fig. 6: (a) skin surface view; (b) sectional view of pigmented nevus.

In the measurement of the attenuation coefficient, this study is divided into three layers from the skin surface downward, the first layer is 1-50 pixels, the second layer is 51-150 pixels, and the third layer is 151-250 pixels, and the fitting results are shown in Figure 9; In order to make the fitting error smaller, both curves are averaged from 300 adjacent A-scans (axial scans) in the region of interest; it can be seen from Figure 8 that due to the scattering of normal skin and pigmented nevus The characteristics are different. In the first 150 pixels, the normal skin tissue has a higher A-scan gray value than the diseased tissue. The first reflection peak of the A-scan of the normal tissue appears later than the A-scan of the diseased tissue. About 25 pixels behind, the attenuation coefficient of pigmented moles in the first layer and the second layer is higher than that of normal skin. In the third layer, the A-scan curves of the two almost overlap. The specific experimental results are shown in Table 3;

In the measurement of fractal dimension, a total of 20 B-scans separated by 5 are selected for analysis, and the area size is 64×64 pixels. The results are shown in Figure 10 and Table 4. It can be seen from Table 4 that the fractal dimension of the pigmented mole area The difference between the number and normal skin tissue was obvious (P<0.05), and the fractal dimension value increased by 3.56%, which indicated that the cells in the pigmented nevus area were more disordered and irregular, and the internal composition was more complex.

Figure 7 A B-scan of the skin and the B-scan after flattening;

In Figure 7: (a) B-scan acquisition image of the skin, the raised area is the display of the pigmented nevus area on this B-scan; (b) The B-scan image based on the flattened skin surface, the dashed frame on the left is normal skin The A-scan region of interest of the nevus is shown in the dashed box on the right.

In Figure 10: the area of interest in normal skin tissue is in the upper dashed box, and the pigmented mole is in the lower dashed box.

table 3

Table 4

To sum up, the experimental results obtained in this experiment prove that the dual-modal system can be used to obtain more information about pigmented nevus and identify the lesions of pigmented nevus, and achieve better results.

The basic principles and main features of the present invention and the advantages of the present invention have been shown and described above. Those skilled in the art should understand that the present invention is not limited by the above-mentioned embodiments, and the descriptions in the above-mentioned embodiments and the description are only to illustrate the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have Various changes and modifications fall within the scope of the claimed invention. The claimed scope of the present invention is defined by the appended claims and their equivalents.

Claims (10)

1. a skin imaging system, is characterized in that, comprises OCT subsystem, HSI subsystem and image processing unit; The OCT subsystem: used to emit light signals, and make the light signals form reference light and sample light in the skin area through reflection, and at the same time cause the sample light and reference light to interfere to form interference spectral information; The HSI subsystem: used to emit white light signals to the skin area, and to split and condense the reflected light generated by the white light signals irradiating the skin area to obtain spectrograms in all wavelength ranges; The image processing unit: used to generate a depth image of the skin area according to the interference spectrum information, reconstruct and correct the spectrogram to obtain a hyperspectral image of the skin area, and fuse the depth image and the hyperspectral image to obtain a skin image . 2. The skin imaging system according to claim 1, wherein the OCT subsystem comprises: Sweep frequency light source: used to emit laser signals; The first fiber coupler: used to split the laser signal output by the frequency sweep light source, and make the split optical signal enter the reference arm and the sample arm respectively; Reference arm: used to reflect the received incident light to generate reference light; Sample arm: used to emit the received incident light to the skin area to be imaged and form sample light after reflection; The second fiber coupler: used to interfere the sample light and the reference light to form interference spectral information; Balanced detector: used to detect the interference spectrum information and output to the image processing unit. 3. The skin imaging system according to claim 2, wherein the sample arm comprises a scanning galvanometer and a first achromatic lens, and the incident light entering the sample arm is commutated by the scanning galvanometer, and then passes through the first achromatic lens. An achromatic lens focuses on the skin area. 4 . The skin imaging system according to claim 3 , wherein the scanning galvanometer is configured with a high-speed analog output device, and the high-speed analog output device can output an analog signal to control the scanning galvanometer to scan the skin area. 5 . 5 . The skin imaging system according to claim 3 , wherein the front end of the scanning galvanometer is further configured with a first optical fiber collimator, which is used for collimating the light signal incident on the scanning galvanometer. 6 . 6. The skin imaging system according to claim 2, wherein the reference arm comprises a second optical fiber collimator, a third optical fiber collimator, a first plane mirror and a second plane mirror, and the incident light entering the reference arm The light is collimated by the second optical fiber collimator and then reflected by the first plane mirror and the second plane mirror and then enters the third optical fiber collimator. 7. The skin imaging system of claim 1, wherein the HSI subsystem comprises: White light source: used to emit white light signals to the skin area; Front optical system: used to achromatize the reflected light generated by the white light signal irradiating the skin area, and make the reflected light pass through the preset incident slit through focusing processing; Spectrometer: used for push-brooming according to a preset step size, and performing spectroscopic and converging processing on the reflected light passing through the incident slit to obtain the spectrum in all wavelength ranges of the skin area; wherein, the step size is equal to the width of the incident slit ; CCD camera: used to detect the spectrum output by the spectrometer to form a spectrogram. 8 . The skin imaging system according to claim 7 , wherein the spectrometer is configured with a motorized translation stage for driving the spectrometer to run to complete the push-broom operation. 9 . 9 . The skin imaging system according to claim 7 , wherein the front optical system comprises at least two second achromatic lenses arranged coaxially. 10 . 10. The skin imaging system according to claim 1, further comprising a dichroic mirror, the dichroic mirror is used to focus the optical signals emitted by the OCT subsystem and the HSI subsystem on the skin area, and simultaneously Used to reflect the light signal reflected from the skin area to the OCT subsystem and the HSI subsystem, respectively.

Images