Application of a maximum likelihood algorithm to ultrasound modulated optical tomography

Abstract

In pulsed ultrasound modulated optical tomography (USMOT), as an ultrasound (U/S) pulse propagates, it performs as a scanning probe within the sample, and modulates the scattered light spatially distributed along the axis of propagation. By detecting and processing the modulated signal, the information along the U/S axis of the sample (1D image) is studied. The signal is modelled as a convolution of the U/S pulse and the ultrasonic and optical properties of the medium along the U/S focus. Based upon this model, a Maximum Likelihood (ML) method for image reconstruction is established. The ML data inversion technique for a pulsed USMOT signal is, for the first time to our knowledge, investigated both theoretically and practically. The ML method inverts the pulsed USMOT signal to retrieve the spatially varying properties of the sample along the U/S scanning column, and then the optical absorption property can be acquired. The results show that the ML method can serve as a useful fitting tool for a pulsed USMOT signal even in the presence of noise. The work illustrates the application of this iterative algorithm on simulated and experimental data. Experimental results using 5cm thick animal tissue phantoms (scattering coefficient μs is 6.5cm-1) demonstrate that the resolution is better than 100μm using a 10MHz transducer.