Artificial neural network-based determination of denoised optical properties in double integrating spheres measurement
Abstract
Accurate determination of the optical properties of biological tissues enables quantitative understanding of light propagation in these tissues for optical diagnosis and treatment applications. The absorption () and scattering () coefficients of biological tissues are inversely analyzed from their diffuse reflectance (R) and total transmittance (T), which are measured using a double integrating spheres (DIS) system. The inversion algorithms, for example, inverse adding doubling method and inverse Monte Carlo method, are sensitive to noise signals during the DIS measurements, resulting in reduced accuracy during determination. In this study, we propose an artificial neural network (ANN) to estimate and at a target wavelength from the R and T spectra measured via the DIS to reduce noise in the optical properties. Approximate models of the optical properties and Monte Carlo calculations that simulated the DIS measurements were used to generate spectral datasets comprising , , R and T. Measurement noise signals were added to R and T, and the ANN model was then trained using the noise-added datasets. Numerical results showed that the trained ANN model reduced the effects of noise in and estimation. Experimental verification indicated noise-reduced estimation from the R and T values measured by the DIS with a small number of scans on average, resulting in measurement time reduction. The results demonstrated the noise robustness of the proposed ANN-based method for optical properties determination and will contribute to shorter DIS measurement times, thus reducing changes in the optical properties due to desiccation of the samples.
1. Introduction
A quantitative understanding of light propagation in biological tissues is essential in the field of biomedical optics for diagnostic and therapeutic applications.1 Light propagation in biological tissues is mainly characterized by two optical parameters: the absorption coefficient and the scattering coefficient .2 The values of and are determined using a combination of spectroscopic measurements and inversion algorithms.3 To date, several methods have been developed to determine the and characteristics of various bio-tissues.4 In the spectroscopic measurements, a double integrating sphere (DIS) system allows the diffuse reflectance R and the total transmittance T to be measured simultaneously at the same point in a sample with high accuracy.5,6 The combination of the DIS system with inversion algorithms is regarded as the gold standard for measurement of these optical properties.7
Several inversion techniques for and from the R and T characteristics measured by the DIS system have been established; examples include inverse adding doubling (IAD),8 the inverse Monte Carlo (IMC) method,9 lookup tables (LTs),10 and artificial neural network (ANN)11,12 based methods. The IAD method, which is based on numerical solution of the radiation transport equation, offers high-speed inversion.13 The IMC method is based on an iterative calculation performed with the Monte Carlo (MC) model of light transport in tissues and calculates the optical properties accurately for a wide range of absorption and scattering coefficients.14 In the LT and ANN methods, the iterative processes required in the IAD and IMC methods are avoided and the calculation time can then be reduced significantly.10,11,12 Although these methods have been used previously, the estimated and values are affected by the measurement noise signals of both R and T.11
The conventional approaches estimate and values from measured R and T for each wavelength. The measured R and T values include random noise signals in the DIS measurements. These measurement noise signals thus affect the estimation of both and and reduce the accuracy of their determination. This measurement noise is unavoidable, and the effect increases as the measurement time for R and T decreases. Longer measurement times often lead to issues due to changes in the dryness of the bio-tissue, which lead to changes in the optical properties in the DIS measurement. To ensure that the and values are determined accurately, a noise-resistant inversion technique is required.
In this work, we propose an ANN model with multiple wavelength inputs (mwANN) to determine the and values of tissues with high noise robustness. The mwANN is based on the fact that the and spectra are theoretically defined as continuous functions in wavelength.2 The measurement noise is reduced by using the continuity of the and values between the target and neighboring wavelengths. The mwANN accepts the measured spectra of R and T for the target and neighboring wavelengths and outputs noise-reduced and values at the target wavelength. To date, several ANN models have been demonstrated to provide a fast technique for determination of and and consideration of the wavelength-dependent anisotropy factor g and the refractive index n.12 In addition to these functions, the mwANN enables noise-reduced estimation of both and . This improvement to the inversion process can shorten the DIS measurement time because of its high noise robustness, leading to measurements that can reduce the effects of sample desiccation and degeneration over time. This offers an advantage in providing accurate estimation without the need for equipment improvements for reducing the noise levels.
2. Materials and Methods
2.1. Neural network model
Figure 1 shows a schematic illustration of the mwANN that is proposed to determine the noise-reduced and values at a target wavelength. The measured R and T values include measurement noise signals that depend on the DIS system and the measurement conditions. The measured R and T values at the target wavelength and several neighboring wavelengths are input into the mwANN model. The mwANN model reduces the estimation noise for both and at the target wavelength by learning the relevance and the continuity of these spectra. Estimated and spectra were obtained by varying the target wavelength.
The mwANN model, which has three hidden layers that incorporate batch normalization and a rectified linear unit (ReLU) function for activation, was then prepared. The mwANN model was tuned independently according to the noise conditions of the DIS system. The codes were implemented in Python and TensorFlow with the Keras module was used as the deep learning toolkit. Training was performed using Adam optimization. The mean squared error was then minimized as a loss function.
2.2. Dataset preparation
To construct a dataset, and spectra were generated using an approximation model and expressed as a function of the wavelength [nm]. Several spectra for biological tissues, denoted by , can be modeled using the spectra of oxyhemoglobin (), deoxyhemoglobin (), water () and melanin (), as shown below2:
Parameter | Min | Max |
---|---|---|
0.001 | 0.04 | |
0.5 | 1 | |
0.5 | 0.8 | |
0 | 0.04 | |
a | 1 | 1.2 |
b | 2 | 4 |
c | 2 | 80 |
d | 0.15 | 0.7 |
To simulate the measurement noise signals, noise signals based on the assumption of DIS measurement were added to the interpolated and values. and are obtained via three optical measurements, comprising a white light reference measurement, a dark reference measurement and a sample measurement, using the DIS system. The dark reference measurements for or were taken experimentally 20 times over the wavelength range of the dataset to obtain the variance at each wavelength as the detection noise signals. The detection noise distributions were assumed to be Gaussian noise. The noise variance of , denoted by , and that of , denoted by , were calculated via error propagation using the variances of the detection noise signals. Gaussian noise signals with variances of and were generated randomly and then added to the interpolated and values. Therefore, a database consisting of 15,000 sets of , , noise-added and noise-added values was constructed. Several databases were prepared by changing and during the evaluation of the noise robustness.
For a training dataset, target wavelengths were assigned randomly for each set in the constructed database. The and values at the assigned target wavelengths and the continuous noise-added and sequences with the same number of wavelengths as the mwANN inputs were then extracted. The training dataset, which consisted of 15,000 pairs of inputs and outputs from the mwANN shown in Fig. 1, was thus constructed. For the test dataset, 1,000 sets of , , noise-added and values were generated using the same procedure that was used for the database for the training dataset.
2.3. Accuracy evaluation
To evaluate the noise reduction performance, the signal-to-noise ratio (SNR) was used. The SNRs of the estimated and values were calculated independently using the following equation23 :
Comparisons of the noise robustness of the proposed mwANN to the IMC method, a one-wavelength ANN (owANN) and a combination of a Gaussian filter (GF) and the owANN were performed. In the IMC method, and were determined using the methodology presented in the literature.24 In the IMC method, n and g were assumed to be constant at all wavelengths and were set at 1.385 and 0.9. In the owANN, 12,000 sets of a meshed training dataset of , , R and T were used for training, as described in the literature.11 To have the same ranges for the optical properties that were used as the training dataset for the mwANN, the ranges of and were 0–3 mm and 0–120 mm, respectively, and the intervals for and were 0.03 mm and 1 mm, respectively. For the combination of the GF and owANN, R and T values that had been denoised using a GF were input to the owANN, and the values of and were estimated. Within the parameter ranges, the and values were divided into each interval size and all possible different combinations were then generated.
2.4. Sample measurement using DIS system
To measure and , a DIS system was used. Our DIS experimental setup was previously described in detail.24 Briefly, a xenon lamp (L2273 and C8849, Hamamatsu Photonics, Japan) was used as the white light source. The light was focused onto the sample surface into a 1mm diameter spot. The sample was placed between the reflectance and transmittance spheres (CSTM-3P-GPS-033SL, Labsphere). The integrating spheres had ports with a diameter of 10mm. The diffusely reflected and transmitted light was detected through an optical fiber (CUSTOM-PATCH-2243142, Ocean Optics) connected to a spectrometer (MAYP10161, Maya2000-Pro, Ocean Optics).
Chicken breast tissue was prepared for the measurements. Frozen chicken breast tissue was cut into thin slices with dimensions of approximately 1cm × 1cm to cover the ports of the integrating spheres. The thickness of the chicken breast tissue was 2.0mm. The sample thickness was measured at three points using a micrometer (MDE-25MX, Mitutoyo, Japan) and averaged. The sample was then sandwiched between two glass slides using spacers to minimize any change in the optical properties due to compression. The tissue was then sealed with glue to prevent desiccation from occurring during the measurements. The above procedures had been completed within 10min and the measurement was performed immediately after sample preparation. To compare the performance of IMC and the mwANN, a single sample was prepared. The exposure time was set at 500ms and the number of scans averaged in the measurement was set at 1 or 100.
3. Results
3.1. Noise reduction in determination of absorption and scattering coefficients
To demonstrate the noise reduction, the and values were estimated by the trained mwANN, which accepted 31 wavelengths from the simulated and spectra measured using the DIS system. Initially, and spectra generated in the same manner shown in Sec. 2.2 were defined as the true values. From the generated and spectra, the simulated and spectra were obtained. Figure 2(a) shows a set of the simulated and spectra. The and values were determined to be and , respectively. Figure 2(b) shows the and spectra calculated using the IMC and mwANN methods. When compared with the and spectra obtained by the IMC method, the results from the mwANN method showed good agreement with the true values. The spectra estimated by the IMC method showed less accuracy, particularly at wavelengths around 400nm, where light absorption by the hemoglobin was notably strong. This led to the small values that were determined for both and within this wavelength range.25 In contrast, the spectra estimated via the mwANN method showed high accuracy with almost no dispersions. Figures 3(a) and 3(b) show the estimation errors for the IMC and mwANN methods, respectively. The average estimation errors for the and spectra were 0.013mm and 1.59mm for the IMC method and 0.010mm and 0.48mm for the mwANN method, respectively. The accuracy of the optical properties estimated from the noised R and T values was improved.
3.2. Evaluation of the number of input wavelengths
To evaluate the relationships between the SNR and the number of input wavelengths, simulated measured and spectra were prepared by assuming that the number of scans to be averaged in the DIS measurement was 1 or 100. For the number of scans to be averaged of 1, and were and , respectively. For the number of scans to be averaged of 100, and were and , respectively. Figure 4 illustrates the relationship between the SNR and the number of input wavelengths. When the neighboring wavelengths were not taken into account (i.e., one wavelength was input), the SNR was low for both measurement conditions, indicating that the noise reduction procedure had not functioned. In contrast, as the number of input wavelengths increased, the SNR was observed to improve for both and . This improvement occurred because use of a greater number of input wavelengths meant that more spectral shape changes could be trained and used for the estimation process. The use of multiple wavelengths as inputs is useful as an approach to noise reduction for the optical properties. For numbers of input wavelengths greater than 11, the SNRs of the and values estimated from the measurement with one scan exceeded the SNR obtained when 100 scans were averaged without consideration of the neighboring wavelengths (i.e., when one wavelength was input), thus indicating the feasibility of reducing the measurement time by applying the DIS.
3.3. Evaluation of noise robustness
To evaluate the noise robustness of the proposed method, the estimation accuracies obtained for each noise level when using the IMC method, the owANN method, the GF method and the mwANN method with 51 input wavelengths were compared. Figure 5(a) shows the relationship between the SNR and the noise level . was defined as . The mwANN method showed a higher SNR and better estimation accuracy when compared with the IMC, owANN and GF methods, except in the case where was near zero. The SNR, which tended to decrease with increasing noise levels, was improved by using the multi-wavelength model, and the high SNR was retained. The IMC and owANN methods both showed a significant reduction in their SNR characteristics, indicating reduced noise robustness. The mwANN method was more robust to the measurement noise than the other methods. Even at high noise levels, the mwANN method was able to reduce the noise and determine the optical properties accurately. Figure 5(b) shows the coefficient of determination () characteristics with the varying noise levels. The mwANN method kept the at the same level (more than 0.99) as the estimated results from noise-free data () with the IMC method, even at high noise levels. The difference between the true and estimated values was maintained at the same level as the noiseless case, indicating that the mwANN method reduces noise without smoothing out the information of the actual and .
3.4. Experimental verification
The and spectra of the chicken breast tissue were estimated from experimental and characteristics measured using the DIS system for experimental verification. Figure 6(a) shows the and spectra of the chicken breast tissues when measured with the number of scans to be averaged of 1 or 100. The acquisition time for the and spectra for one scan can be shortened to 1/100 of that for 100 scans. Figure 6(b) shows the and spectra obtained via the IMC method when using measured data that were averaged over 100 scans. A strong absorption peak was observed at approximately 412nm and smaller peaks were observed at approximately 548nm and 575nm. These peaks were considered to be derived from hemoglobin. Some blood vessels might be included in the chicken breast tissue sample. The spectrum decreased monotonically with increasing wavelength, which was attributed to a reduction in the Rayleigh scattering contribution and an increase in Mie scattering. Figures 6(c) and 6(d) show the and spectra determined from and when measured with one scan by the IMC and mwANN methods, respectively. The mwANN method reduced the effects of noise propagation on the measurement of both and . The resulting and spectra showed values that were closer to the spectra shown in Fig. 6(b). The small absorption peaks produced by hemoglobin at approximately 548nm and 575nm were identified more clearly by the mwANN method than by the IMC method. These results indicate that the mwANN model can reduce the number of scans required for the DIS measurements, and thus contributes to reduction of the measurement time required to perform spectroscopy.
3.5. Adaptability evaluation
To evaluate the adaptability of the trained mwANN, the and spectra of human subcutaneous fat were estimated based on previously reported and results.24 In the previous work,24 the spectrum has a peak for bilirubin at 475nm, i.e., a spectrum with a shape that differs from that for hemoglobin absorption. Figure 7 shows the and spectra obtained by the IMC and mwANN methods. In the spectrum that was analyzed using the IMC method, the absorption peaks of hemoglobin were observed at wavelengths of approximately 415nm, 450nm and 574nm, while the peak for bilirubin was observed at 475nm. Although spectra with bilirubin absorption bands were not included among the training datasets, the mwANN method was still able to estimate the absorption peak. The spectrum also showed agreement between the IMC and mwANN methods. The peaks of other absorbers could be estimated because the spectral shape changes in R and T were trained appropriately.
4. Discussion
An ANN model that used R and T values at multiple wavelengths as inputs was constructed for noise-reduced measurement of the and spectra. The numerical and experimental results demonstrated that the mwANN was robust to measurement noise and that it improved the estimation accuracy of and from the noised and characteristics. The use of multiple wavelengths as inputs is an effective approach for noise reduction and can be applied regardless of the actual noise level. In addition, accurate and spectra can be calculated from and when measured using a small number of scans in the DIS system, which contributes to reduced DIS measurement times. There are few restrictions with regard to the biological tissues to which the mwANN can be applied because absorption peaks that are not considered in the optical properties model for the dataset can be estimated.
In the mwANN method, noise reduction was achieved by focusing on the nature of the continuity of the absorption and scattering spectra of the biological tissues. The proposed mwANN method was applied over the visible to near-infrared wavelength range in this study. The same algorithm can be used to estimate and over other wavelength ranges, as long as an appropriate training dataset is prepared. In the construction of the training dataset, the main optical absorbers in biological tissues (i.e., hemoglobin, oxyhemoglobin, water and melanin) were considered. Because the optical properties were estimated from spectral changes in and , absorption bands that were not considered as part of the training data were also detected, as shown in Fig. 7. Other absorption peaks may be detected, along with that of bilirubin.
The results from the owANN and mwANN methods were compared as shown in Fig. 5. When using the owANN method, the values of the estimated and were 0.9684 and 0.9905, respectively, for the simulated and when assuming DIS measurements with 100 scans. The mwANN method estimated the and spectra more accurately, with values of 0.9944 and 0.9932, respectively. The mwANN method determined and with higher values than the owANN method, even in measurements performed with measurement noise variance that was 60 times larger. The mwANN model was more robust to noise signals when compared with the owANN model. The mwANN can shorten the acquisition time by using measured data that have been averaged over smaller numbers of scans in the DIS system. At least three measurements are required for or in conventional DIS measurements, making the measurement time relatively long.26 Prolonged measurements might lead to changes in the optical properties of the biological sample. Zhu et al. reported that the absorption coefficient of porcine liver tissue increased over time because of the weight loss and water loss caused by drying.27 By using the mwANN method, it is expected that the reduced DIS measurement time requirement will enable measurements that minimize any changes in the optical properties caused by the sample conditions.
Although the input to the mwANN will need to be modified, an ANN-based approach with multiple wavelength inputs could be implemented for in vivo measurements when using broadband wavelengths. In diffuse reflection spectroscopy (DRS), the absorption and scattering coefficient spectra were estimated by the IMC28 and ANN29 methods from spatially resolved diffuse reflectance. Generally, the SNR in DRS decreases with increasing source-detector separation, or with increasing distance between the optical fibers that deliver and collect the light.30 The potential for noise reduction when using an ANN-based approach may provide a higher SNR for in vivo optical properties. Furthermore, the time required to perform in vivo spectroscopy can be reduced, thus contributing to real-time measurement of the optical properties.
5. Conclusions
This study has presented an ANN-based method for noise-resistant determination of the optical properties of biological tissues. The ANN was trained to output and values at a target wavelength based on the inputs of R and T at the surrounding wavelengths. The ANN-based method demonstrated better noise reduction performance than the conventional methods in this field. Furthermore, in vitro experiments showed that the ANN can contribute to the performance of short-time DIS measurements and can be applied to samples with absorption peaks that are not included in the dataset. This method will improve the accuracy of optical properties determination and will enable DIS measurements with minimized effects from changes in the optical properties over time.
Acknowledgments
This work was supported by the Japan Society for the Promotion of Science KAKENHI (Grant numbers: 20H04549 and 19K12822) and the Japan Science and Technology Agency ACT–X (Grant Number: JPMJAX21K7).
Conflicts of Interest
The authors have no conflicts of interest relevant to this article.
References
- 1. , “Combination of diffuse reflectance and transmittance spectroscopy to obtain optical properties of liquid phantoms,” Opt. Eng. 59(2), 024109 (2020). Crossref, Web of Science, Google Scholar
- 2. , “Optical properties of biological tissues: A review,” Phys. Med. Biol. 58(11), R37–R61 (2013). Crossref, Web of Science, Google Scholar
- 3. , “Optical properties of skin, subcutaneous, and muscle tissues: A review,” J. Innovative Opt. Health Sci. 4(1), 9–38 (2011). Link, Web of Science, Google Scholar
- 4. , “A review of in-vivo optical properties of human tissues and its impact on PDT,” J. Biophotonics 4(11–12), 773–787 (2011). Crossref, Web of Science, Google Scholar
- 5. , “Double-integrating-sphere system for measuring the optical properties of tissue,” Appl. Opt. 32(4), 399–410 (1993). Crossref, Web of Science, Google Scholar
- 6. , “Optical properties of human skin in the near infrared wavelength range of 1000 to 2200nm,” J. Biomed. Opt. 6(2), 167–176 (2001). Crossref, Web of Science, Google Scholar
- 7. , “Biomedical applications of integrating sphere: A review,” Photodiagnosis Photodyn. Ther. 31, 101712 (2020). Crossref, Web of Science, Google Scholar
- 8. , “Determining the optical properties of turbid media by using the adding–doubling method,” Appl. Opt. 32(4), 559–568 (1993). Crossref, Web of Science, Google Scholar
- 9. , “Optical properties of ocular fundus tissues-an in vitro study using the double-integrating-sphere technique and inverse Monte Carlo simulation,” Phys. Med. Biol. 40(6), 963–978 (1995). Crossref, Web of Science, Google Scholar
- 10. , “Lookup table–based inverse model for determining optical properties of turbid media,” J. Biomed. Opt. 13(5), 050501 (2008). Crossref, Web of Science, Google Scholar
- 11. , “Artificial neural network method for determining optical properties from double-integrating-spheres measurements,” Chin. Opt. Lett. 8(2), 173–176 (2010). Crossref, Web of Science, Google Scholar
- 12. , “Determination of optical properties in double integrating sphere measurement by artificial neural network based method,” Opt. Rev. 28(1), 42–47 (2021). Crossref, Web of Science, Google Scholar
- 13. , “In vitro determination of normal and neoplastic human brain tissue optical properties using inverse adding-doubling,” Phys. Med. Biol. 51(8), 2011–2027 (2006). Crossref, Web of Science, Google Scholar
- 14. , “Monte Carlo-based inverse model for calculating tissue optical properties. Part I: Theory and validation on synthetic phantoms,” Appl. Opt. 45(5), 1062–1071 (2006). Crossref, Web of Science, Google Scholar
- 15. , “Rapid and accurate estimation of blood saturation, melanin content, and epidermis thickness from spectral diffuse reflectance,” Appl. Opt. 49(10), 1707–1719 (2010). Crossref, Web of Science, Google Scholar
- 16. , “Measurement of tissue optical properties in the context of tissue optical clearing,” J. Biomed. Opt. 23(9), 091416 (2018). Crossref, Web of Science, Google Scholar
- 17. , “Tomographic bioluminescence imaging by use of a combined optical-pet (OPET) system: A computer simulation feasibility study,” Phys. Med. Biol. 50(17), 4225–4241 (2005). Crossref, Web of Science, Google Scholar
- 18. , “Quantification of the optical properties of two-layered turbid media by simultaneously analyzing the spectral and spatial information of steady-state diffuse reflectance spectroscopy,” Biomed. Opt. Exp. 2(4), 901–914 (2011). Crossref, Web of Science, Google Scholar
- 19. , “MCML-Monte Carlo modeling of light transport in multi-layered tissues,” Comput. Methods Progr. Biomed. 47(2), 131–146 (1995). Crossref, Web of Science, Google Scholar
- 20. , “Next-generation acceleration and code optimization for light transport in turbid media using GPUs,” Biomed. Opt. Exp. 1(2), 658–675 (2010). Crossref, Web of Science, Google Scholar
- 21. , Tissue Optics: Light Scattering Methods and Instruments for Medical Diagnostics, Third Edition, SPIE, Bellingham, Washington (2007). Crossref, Google Scholar
- 22. , “In vitro double-integrating-sphere optical properties of tissues between 630 and 1064nm,” Phys. Med. Biol. 42(11), 2255–2261 (1997). Crossref, Web of Science, Google Scholar
- 23. , “Motion artifact cancellation in NIR spectroscopy using discrete Kalman filtering,” Biomed. Eng. Online 9(1), 1–10 (2010). Crossref, Google Scholar
- 24. , “Measurement of absorption and reduced scattering coefficients in Asian human epidermis, dermis, and subcutaneous fat tissues in the 400-to 1100-nm wavelength range for optical penetration depth and energy deposition analysis,” J. Biomed. Opt. 25(4), 045002 (2020). Crossref, Web of Science, Google Scholar
- 25. , “Optical properties of human skin, subcutaneous and mucous tissues in the wavelength range from 400 to 2000nm,” J. Phys. D Appl. Phys. 38(15), 2543–2555 (2005). Crossref, Web of Science, Google Scholar
- 26. , “Double-integrating-sphere system at the National Institute of Standards and Technology in support of measurement standards for the determination of optical properties of tissue-mimicking phantoms,” J. Biomed. Opt. 20(12), 121310 (2015). Crossref, Web of Science, Google Scholar
- 27. , “Effects of dehydration on the optical properties of in vitro porcine liver,” Lasers Surg. Med. 33(4), 226–231 (2003). Crossref, Web of Science, Google Scholar
- 28. , “Inverse Monte Carlo method in a multilayered tissue model for diffuse reflectance spectroscopy,” J. Biomed. Opt. 17(4), 047004 (2012). Crossref, Web of Science, Google Scholar
- 29. , “Efficient construction of robust artificial neural networks for accurate determination of superficial sample optical properties,” Biomed. Opt. Exp. 6(3), 747–760 (2015). Crossref, Web of Science, Google Scholar
- 30. , “Sampling depth of a diffuse reflectance spectroscopy probe for in-vivo physiological quantification of murine subcutaneous tumor allografts,” J. Biomed. Opt. 23(8), 085006 (2018). Crossref, Web of Science, Google Scholar