Quantifying floral shape in precision is crucial for studying the morphological variation of flowers. This work aims to measure the floral shape of Sinningia speciosa (Darwin’s Gloxinia) by using two-dimensional (2D) and three-dimensional (3D) floral images, and to examine the shape variation by using geometric morphometrics (GM). S. speciosa has two shape forms – bilateral symmetric wild type and radially symmetric peloria. In this work, the wild-type S. speciosa was crossed with the peloria to create the floral shape polymorphism. 2D and 3D floral images of second generation (F2) were obtained by using digital camera and micro-computed tomography (micro-CT), respectively. Their floral shapes were represented by a set of characteristic landmarks. GM was then applied to the landmarks to assess floral shape variations. ANOVA demonstrated the association between the floral shape variations and the SsCYC genotypes. It was shown that the proposed approach can adequately capture the shape variation of F2 flowers.