Modern surgical navigation technologies like AI-assisted systems rely heavily on 3D models of human organs to plan and guide procedures. However, the availability of these annotated datasets is very limited due to privacy rules and the complexity of medical data. This project tackles that problem by presenting a pipeline that can create new, realistic combinations of abdominal organs using real 3D scans from existing public datasets. These new sets can then be used to train and test AI systems without needing more real patient data. To make sure the results still resemble real anatomy, the project uses different metrics to check the size, shape, and position of the organs and evaluates how well the newly created combinations reflect real human variation. This method helps increase the diversity of training data for medical AI and supports the development of safer, smarter surgical tools.