The ALISA Component Module (ACM) has been developed as a general-purpose shape classifier for objects in Digital Transmissive Images (DTIs). The Component Module classifies contoured DTIs intended to reveal the internal structures of objects and to facilitate feature extraction. The ACM is trained with exemplars of regions-of-interest (ROIs) representing object components.??While the ACM has demonstrated robust classification performance for a few application domains, its performance as a general shape classifier remains to be determined. Therefore, a set of general shapes is needed to measure the ACM's performance as a general shape classifier. This paper presents a novel approach to generating such a set of shapes by contouring randomly generated fractal images and selecting subsets of contours analogous to ROIs of objects in DTIs. Linear boundary regression (LBR) is postulated as a post-process to accommodate the generation of contoured ROIs more closely resembling contours of human-made objects in DTIs.