Optical image reconstruction based on the third-order diffusion equations

This paper presents a third-order diffusion equations-based optical image reconstruction algorithm. The algorithm has been implemented using finite element discretizations coupled with a hybrid regularization that combines both Marquardt and Tikhonov schemes. Numerical examples are used to compare between the third- and first-order reconstructions. The results show that the third-order reconstruction codes are more stable than the first-order codes, and are capable of reconstructing void-like regions. From the examples given, it has also been shown that the first-order codes fail to both qualitatively and quantitatively reconstruct the void-like regions.