Functional diffuse optical tomography (fDOT) system for breast imaging consists of an optical fiber-based light delivery subsystem to sequentially inject multi-wavelength NIR light at multiple locations on the tissue surface and a diffuse transmitted light measurement subsystem. In the low-cost fDOT, few NIR light sources and detectors are electromechanically multiplexed. Though it reduces cost substantially, it adds measurement uncertainty and this increases with the measurement cycle. Traditionally homogeneous phantom measurement data is used for the high-level system calibration. In this paper, an embedded system based digital calibration technique in the hardware level for the electromechanical optical fiber switch-based 3D fDOT system is proposed that is suitable for the low-resource settings. The system has four LED sources of four wavelengths (660, 735, 810 and 850 nm) and 24 SiPD detectors. An algorithm was developed and programmed a microcontroller-based circuitry to digitally control the electromechanics with a spatial resolution of 12.5 micro meter to couple the four-wavelength NIR sources to sixteen source fibers. A calibration scheme was adopted for source illuminations that takes feedback from a power meter to the controller and digitally calibrates to ensure that light entering the imaging domain is near identical. Measurement data from a homogeneous and a heterogeneous phantom was used to study measurement uncertainty and noise performance, and to compare with the traditional method. Experimental results statistically showed that active hardware-level digital calibration improved the measurement accuracy and convergence of the image reconstruction that can open up to a fast, reliable, and cost-effective fDOT system.
|