A computer-aided diagnosis system that can identify the different human tissue and organ regions from CT images has been developing in our research group. In this paper, we describe the details of the extraction procedures for skin, soft tissue and skeleton regions that are the first step of the recognition process. The relationship between the different human organs and the density features of each organ region has been used in the identification process. We applied this process to 8 patient cases of multi-slice CT images; each one covers the whole torso of human body with the isotropic resolution of about 0.63 mm voxel. The experimental results showed that the procedures could recognize the target regions from the original CT images.