Estimation of partial volume effect using spatial context. Application to morphometry in cerebral imaging [MRI]

The purpose of the work presented in this article is the estimation of the partial volume effect very often observed in medical image processing. The authors' working hypothesis is that such partial volume effects will occur in tissue interfaces, the proposed method uses a tissue-labeled distance map to detect tissue interfaces as being those pixels closer to two particular tissues than to any others. The tissue labeled distance map is also used to detect pure tissue means as being the gray-level means of pixels far from tissue interfaces. Knowing the positions of the different interfaces as well as the pure tissue means, the authors are able to compute the proportion of the two tissues mixed in each pixel of the interface.