Mr. Liyu Dai Profile

Liyu Dai

SPIE Involvement:

Author

Publications (1)

Proceedings Article | 10 October 2020 Presentation + Paper

Study of spatial frequency domain imaging technique for turbid media optical property estimation and application

Xiaping Fu, Xu Jiang, Liyu Dai, Yifeng Luo

Proceedings Volume 11552, 115520Q (2020) https://doi.org/10.1117/12.2573928

KEYWORDS: Optical properties, Machine learning, Absorption, Scattering, Biological research, Light scattering, Statistical analysis, Diffuse reflectance spectroscopy, Tissue optics, Tissues

Read Abstract +

The interaction phenomenon between light and turbid media like tissue is mainly characterized of absorption and scattering. The absorption coefficient (μa) and reduced scattering coefficient (μ's) of biological tissue can provide useful information about the composition concentration and structure features. Spatial frequency domain imaging (SFDI) is a relatively new and effective technique for quantitative optical property mapping of turbid media, with unique advantages of noncontact and wide-field detection. In this study, a self-developed SFDI system was used for optical property detection of phantom and meat tissue. System calibration and error correction was realized by using a series of self-made phantoms that covering a wide range of absorption and reduced scattering. Optical properties of turbid media samples of different meat tissue (pork tenderloin, pork hind leg muscle, beef sirloin, beef rump steak, chicken breast and chicken drumstick) were estimated based on spatial frequency dependent diffuse reflectance. The results showed that suitable system calibration and error correction can help to improve the system performance and detection accuracy. It’s found that the optical properties (μa and μ's) of different species of meat were distinctively distinguished. The classification accuracy of using μa and μ's as features was higher than the result of using reflectance as the feature, except for chicken breast and thigh, which indicated the advantage of SFDI in measuring the properties of meat samples compared to traditional spectroscopy. Then the estimated optical properties were furtherly applied for sample classification. Classification accuracies of different meat species and parts were much higher by using SVM compared with using KNN. The classification accuracy was as high as 96.7% for three meat species classification, and 90.8% for six set of meat samples classification. Future studies could be focused on multi spectral imaging of optical properties of meat tissue as well as other biological tissue to estimate related physical or chemical properties quantitatively.

View contact details

UPDATE YOUR PROFILE

Is this your profile? Update it now.

Sign into your SPIE.org account

Don’t have a profile and want one?

Create an account on SPIE.org

Keywords/Phrases

Search In:

Publication Years