This paper presents an automated approach in three-dimensional digital model generation of human joints based on MRI and CT scans. Since both the MRI and CT scans are going to be used in the digitization process, a systematic approach had to be taken in obtaining these scans. In this approach the CT and MRI scans are mapped in order to obtain the exact geometry of human joint. Due to the poor visibility of the hard or soft tissue contours because of the type of imaging used, additional smoothing techniques had to be developed. Attachment points of the ligaments are also identified. As it is well known in design and engineering area that the boundary locations and conditions are important in correct computations of the systems response. The approach described is then applied to a human tibia-femoral joint with soft tissue. The method described in this paper is applicable to live subjects but a cadaveric knee is used here in order to allow verification of results by other means. The image synchronization process which will be described in the paper is used. Non-metal hardware was created to constrain the knee without interfering with signals from the MRI scanner. Then, the tissue boundaries were digitized in each image. In order to obtain a smooth transition from this level to a finite element solid modeler an automated module was developed. The hard tissue, soft tissue and any region with different material properties are kept as a separate zone. The digital model obtained is now ready to be used in the finite element analysis and in reconstructive surgical design of the knee joint. Three dimensional digital images of a human knee joint generated using a cadaveric knee are presented.