3D RECONSTRUCTION AND EVALUATION OF TISSUES BY USING CT, 
MR SLICES AND DIGITAL IMAGES 
S. Dogan 
Ondokuz Mayis University, Engineering Faculty, Dept of Geodesy and Photogrammetry, 55 139 Kurupelit, Samsun, 
Turkey- sedatdo@omu.edu.tr 
Commission V, WG V/3 
KEY WORDS: Three-dimensional, Medicine, Modelling, Photogrammetry, Registration, Photo-realism, Volume Rendering, 
Surface Rendering. 
ABSTRACT 
In this project, it is provided to reconstruct 3D models of human body by using CT, MR slices and digital images and precisely 
finding locations of pathological formations such as tumours. For this purpose, within this project we developed a software, which 
we called as “Medical Image Processing and Analysis System (MIPAS)” and it is still under development for multipurpose medical 
applications. In this paper, we introduce the abilities of MIPAS briefly and also give a sample application on finding location and 
visualization of a brain tumour. MIPAS uses volume and surface rendering techniques for 3D modelling of the tissues and provides 
both volume and surface models at the same time on the screen when required. MIPAS gives many editing and analysis 
functionalities to medical doctors. For pre-processing of CT and MR images, there are wide range of image processing 
functionalities. It is possible to register CT and MR images by using both anatomical landmarks and artificial (external) markers. 
MIPAS also provides surface registration functions with both rigid body and non-rigid body transformation with the variants of ICP 
algorithm. For photo-realistic visualization of the external human body like face of the patient, it provides a photogrammetric 
module. This module consists of self calibration with bundle adjustment by using additional lens parameters, automatic image 
coordinate measurement with ALSM matching technique and texture mapping functions with both colinearity equations and 3D 
affine transformation. We are still studying to progress the photogrammetric module for plastic surgery. Up to now, we could not 
still test the MIPAS on the real patients. But we have just started a new project with medical doctors, in order to test the MIPAS on 
real patients. During this new project, we will see the actual reliability of the system. 
1. INTRODUCTION 
Medical Imaging techniques are used for diagnosing and 
treatment of many diseases as well as surgical operations. CT 
and MR imaging techniques are the mostly used ones. 
Reconstruction of 3D volume and surface models of the tissues, 
by using 2D image slices, provides many advantages to medical 
doctors. For a long time, 3D models have being used in medical 
applications. During the treatment period, tracing the temporal 
changes of the abnormalities is a very important task for 
deciding whether the treatment is positively effective or 
something going wrong. For detection of changes that appear in 
tissues, firstly the location and the geometric quantities of the 
abnormal regions are required. For example, the volume and the 
surface area of the tumour are to be known for temporal 
comparisons. Another example may be given for the treatment 
of Parkinson disease. Furthermore, it is also important to know 
the location of the abnormalities according to some reference 
points which were cited on the external surfaces of the body 
such as human face (Altan and Dogan, 2003). 
When someone searches the scientific literature on above tasks, 
he/she will see that these topics are not directly related to 
photogrammetry. They are especially related to medical and 
biomedical imaging techniques that comprise image processing, 
electronic medicine, radiology and computer graphics. On the 
other hand, one of the primary aims of photogrammetry is 
spatial measurement of objects. Near this, in photogrammetry 
textbooks, some will see the chapters on medical 
photogrammetry or potential use of photogrammetry on medical 
applications. But due to the nature of basic data and 
mathematical basis of photogrammetry, the potential medical 
applications had been limited by the thoughts restricted with the 
applications could only be carried out with perspective 
projection. With this thought in the past, especially in 
orthopedics, photogrammetric techniques had been used with 
the roentgen films based on perspective projection, in order to 
find the 3D location of the pathologies (Elad and Einav, 1996). 
Almost with the same approaches, photogrammetry had also 
been used in orthodontics applications (Schewe et. al, 1999). 
Providing a long bibliography would be easy. However the list 
of cited studies are limited (Mitchell and Newton, 2002). In 
photogrammetry, there are some studies on 3D human body 
measurements. These studies have not been carried out for 
medical purposes, but for art, forensic sciences or various other 
applications which are not directly related to medical purposes 
(Mitchell and Newton, 2002). Up to now, photogrammetric 
contributions were aiming mainly at the reconstruction of the 
outer body using optical sensors, the mapping of the internal 
body presents a much greater research area; both in terms of 
scientific challenge and in terms of the wider range of 
applications  (Patias, 2002). When looking at the 
photogrammetry and medical imaging at the first time, someone 
might think that these are very different topics. Especially, 
when CT, MR or such other radiological imaging modalities are 
considered, someone possibly feel that these are not related 
each other. But when considering the final purposes 3D object 
reconstruction and determination of the spatial location of 
object, it would be clear that these two topics are closely related 
to each other by the means of both mathematical models and 
ds 
TER