Application of the artificial neural network for reconstructing the internal-structure image of a random medium by spatial characteristics of backscattered optical radiation

Journal Article · Mon Jun 30 04:00:00 EDT 2008 · Quantum Electronics (Woodbury, N.Y.)

DOI:https://doi.org/10.1070/QE2008V038N06ABEH013843· OSTI ID:21466875

Veksler, B A ^[1]; Meglinskii, I V ^[2]

---

1. Cranfield Health, Cranfield University, Silsoe (United Kingdom)
2. N. G. Chernyshevskii Saratov State University, Saratov (Russian Federation)

The feasibility of using an artificial neural network (ANN), which is the standard Matlab tool, for non-invasive (based on the data of backscattering) diagnostics of macro-inhomogeneities, localised at subsurface layers of the turbid strongly scattering medium was shown. The spatial and angle distribution of the backscattered optical radiation was calculated by using the Monte-Carlo method combining the modelling of effective optical paths and the use of statistical weights. It was shown that application of the backscattering method together with the ANN allows solving inverse problems for determining the average radius of the scattering particles and for reconstructing the images of structural elements within the medium with a high accuracy. (special issue devoted to application of laser technologies in biophotonics and biomedical studies)

• OSTI ID: 21466875
• Journal Information: Quantum Electronics (Woodbury, N.Y.), Journal Name: Quantum Electronics (Woodbury, N.Y.) Journal Issue: 6 Vol. 38; ISSN 1063-7818
• Country of Publication: United States
• Language: English

Related Subjects

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND TECHNOLOGY
ACCURACY
ANGULAR DISTRIBUTION
BACKSCATTERING
CALCULATION METHODS
COMPUTER CODES
DISTRIBUTION
IMAGES
LASERS
LAYERS
M CODES
MONTE CARLO METHOD
NEURAL NETWORKS
PARTICLES
SCATTERING
SIMULATION
SPATIAL DISTRIBUTION