Regression analysis on forward modeling of diffuse optical tomography system for carcinoma cell detection

Autor: K. Uma Maheswari, M. Thilak, N. SenthilKumar, N. Nagaprasad, Leta Tesfaye Jule, Venkatesh Seenivasan, Krishnaraj Ramaswamy
Jazyk: angličtina
Rok vydání: 2023
Předmět:
Zdroj: Scientific Reports, Vol 13, Iss 1, Pp 1-15 (2023)
Druh dokumentu: article
ISSN: 2045-2322
DOI: 10.1038/s41598-023-29063-4
Popis: Abstract The forward model design was employed in the Diffuse Optical Tomography (DOT) system to determine the optimal photonic flux in soft tissues like the brain and breast. Absorption coefficient (mua), reduced scattering coefficient (mus), and photonic flux (phi) were the parameters subjected to optimization. The Box–Behnken Design (BBD) method of the Response Surface Methodology (RSM) was applied to enhance the Diffuse Optical Tomography experimental system. The DC modulation voltages applied to different laser diodes of 850 nm and 780 nm wavelengths and spacing between the source and detector are the two factors operating on three optimization parameters that predicted the result through two-dimensional tissue image contours. The analysis of the Variance (ANOVA) model developed was substantial (R2 = > 0.954). The experimental results indicate that spacing and wavelength were more influential factors for rebuilding image contour. The position of the tumor in soft tissues is inspired by parameters like absorption coefficient and scattering coefficient, which depend on DC voltages applied to the Laser diode. This regression method predicted the values throughout the studied parameter space and was suitable for enhancement learning of diffuse optical tomography systems. The range of residual error percentage evaluated between experimental and predicted values for mua, mus, and phi was 0.301%, 0.287%, and 0.1%, respectively.
Databáze: Directory of Open Access Journals
Nepřihlášeným uživatelům se plný text nezobrazuje