Autonomous point-based registration of prostate gland tissue images

Three-dimension reconstruction of histological tissues requires accurate image registration. Designing an image registration system strongly relies on the nature of images; manual techniques suffer from a lack of precision and are heavily time consuming. Various automatic or semi-automatic systems exist to address this problem for particular images. In this paper, we provide an entirely autonomous image registration system adapted for prostate tissue slices alignment. This system involves an autonomous way of selecting control points to get precise registration combined with simple image preprocessing to obtain uniform results regardless of the image set variability. We also provide quantitative results based on mutual information and XOR metrics.