The definition of 3D Finite Element (FE) volume models of female torso requires geometry information of naked surface and material properties of human tissues. Analog methods determining breast volume by water immersion of breast or plaster casts are still referred to as gold standard [1]. However, the human body surface is nowadays digitized by 3D scanners, which have become handy and affordable in the last years [2,3]. The term segmentation describes the strategy for separating breast tissue region from 3D surface scans of the female thorax. Different segmentation strategies are proposed in literature, either by outline definition [3,4,5,6] or by creating a parametric surface (Coons patch) from boundary curves using fiducial points [1,4]. The volume between breast base (artificial chest wall) and skin surface forms breast volume. In this study, 59 participants (19 to 67 years, bra size 75B to 95G) were scanned with a low-cost hand-held 1 st generation Sense TM 3D scanner (3D Systems Inc., Rock Hill, SC, USA) in two different positions: standing upright on a turntable and lying on the back (supine), both with the palms of their hands resting on the anterior superior iliac spine. The supine position increases the visibility of the inframammary fold, a common problem especially in ptotic breasts [7]. The breast outline was marked with skin marker. From the 3D scan data in *.ply file format containing geometric and color information, triangular elements representing breast tissue and other regions were selected in open source software Blender 2.79b [8]. All selected regions were exported separately as *.stl files for further data processing in FE pre-processor Patran 2014.1 (MSC.Software Corporation, Santa Ana, CA, USA), where breast base was created and breast volume was calculated. Breast volume was compared to bra size and sister size groups, respectively, which usually shows relatively low accordance [9], indicating the importance of application and market specific determination of breast shape and volume.