Optical properties reconstruction in nonhomogeneous participating medium based on an improved sequential quadratic programming

Abstract

The reconstruction of optical parameters achieves functional imaging by providing characteristics closely related to the physiological state of biological tissues. With considerable research significance, this imaging technology has received increasing attention in the field of medical diagnosis and treatment. However, owing to the limited number of known measurement signals far less than the number of target parameters to be estimated, reconstruction requires a long ill-conditioned process, which is easily polluted by background noise so as to difficulties for achieving high imaging accuracy and efficiency. To conquer these deficiencies, an improved sequential quadratic programming (ISQP) is proposed in this work. In the ISQP, dual constraints of equality and inequality are added in the process of minimizing the objective functional, which avoids the time-consuming forward problem. Moreover, duality method, reduced Hessian matrix, and second-order correction are introduced in the subproblem. In addition, prior information of diagonal contiguous points is introduced to overcome the ill-posedness. The reconstruction performance of ISQP in different modes is discussed in depth. A time-frequency coupling model based on ISQP is constructed subsequently, of which the practical significance is demonstrated in the biological tissue-like medium model. The results show that in the established frequency domain model, ISQP significantly shortens the calculation time of the traditional method and improves the reconstruction quality of absorption parameters. The developed time-frequency coupling method effectively eliminates the influence of the background disturbance, and in the meantime improves the accuracy and efficiency when simultaneously estimating multi-parameters related to optics, with good robustness. The effectiveness of the ISQP combined with hybrid time-frequency model provides new referable ideas and application potential for the development and progress of medical, biology, photonics and other fields.