The creation of highly detailed components presented in this paper builds upon our previously introduced technique for creating complete models from only a small number of measured seed points. Here, we aim at increasing the level of automation in 3D reconstruction of scenes from a small number of images. We build highly detailed model components, for example; windows, columns, groin-vaulted ceilings elements, and arches with large number of automatically created points. The components are created to be reusable in other parts of the model, or any other model, which are similar in shape, but may vary in dimensions. Our approach works directly on the images and performs accurate transformations and scaling automatically without trials and errors. We developed a copy and paste procedure that automatically re-triangulates the base model to account for the pasted element. By creating reusable higher-level building blocks that can be integrated to form a highly detailed model, a growing library of image-based reusable components will materialize. This will significantly reduce the effort, time, and expertise required to create detailed 3D models from images.