The National Cancer Institute has collected a large database of uterine cervix images, termed “cervigrams” for cervical cancer screening research. Tissues of interest within the cervigram, in particular the lesions, are of varying sizes and complex, non-convex shapes. The current work proposes a new methodology that enables the segmentation of non-convex regions, thus providing a major step forward towards cervigram tissue detection and lesion delineation. The framework transitions from pixels to a set of small coherent regions (superpixels), which are grouped bottom-up into larger, non-convex, perceptually similar regions, utilizing a new graph-cut criterion and agglomerative clustering. Superpixels similarity is computed via a combined region and boundary information measure. Results for a set of 120 cervigrams, manually marked by a medical expert, are shown.