Visualization Algorithms for Orthopedic Surgery Simulation

Abstract

During common orthopedic surgery training, students must learn how to perform numerous surgical procedures like fixing fractures which requires training on artificial bones with the usage of surgical tools and implants. These artificial bones have a high cost that depends on the bone's type and quality. Thus the idea of using a computer based simulators for orthopedic surgery training appeared. Simulators will decrease the cost and help students to practice various procedures on a large number of available simulated surgeries in a safe and controlled environment. Also as they are designed dedicatedly for training with specific training goals and high quality visualization of bones, they will have much more to offer than artificial patterns.

Visual representation of the bone is the key element in orthopedic surgerysimulation. There were many sensors used to manage this; such as magnetic resonance imaging, computed tomography, and ultra-sound. The past few decades have witnessed an increasing number of new techniques being developed for medical image visualization, which brought profound changes to personal health programs and clinical healthcare delivery. It seeks revealing internal structures hidden beneath skin and muscles, as well as diagnoses and treatment.

Nowadays smart phones have become an essential part of modern human lives and got involved in so many aspects with a wide variety of applications, not excluded from this; health care professionals who can use this device in one of the most prominent aspects of life, health care. Mobile devices became the trend toward information systems and ubiquitous graphical devices and native volume rendering due to their rapid development in the graphics hardware which can be similar to the PCs. New graphic application programming interfaces (APIs) have been developed as worthy candidates to support the volume rendering core such as; OpenGL ES, Microsoft DirectX 12, AMD’s Mantle, and recently Metal API from apple. With the rare existence of medical visualization applications on mobile devices, this study project light on visualization on mobile system and the recent mobile application.

The thesis presents a survey of the recent intelligent techniques and algorithms used for processing medical data visualization. These techniques cover filtering, segmentation, classification and visualization, with a discussion of each process and comparison between the techniques used in each one. The study shows that direct rendering don’t need to apply high preprocessing techniques before the visualization step whichoffer much greater flexibility when compare to indirect rendering which consideredas standard for nearly all 3D visualization problems. In addition, indirect rendering cannot detect the information inside the object unlike direct rendering. The study also, presents the recent toolkits and software supporting medical volume visualization.

Based on the literature review and dueto the rare existence of free open source medical visualization mobile applications.This study represents mobile applicationthat will help medical students and doctors to show bones processed over in an orthopedic surgery.Experimental results obtained visually by comparing the visualized object from the implemented application to the object obtained from ImageVis3D application. ImageVis3D's development was initiated in 2007 by the NIH/NCRR Center for Integrative Biomedical Computing and additionally supported by the DOE Visualization and Analytics Center for Enabling Technologies at the SCI Institute. The results show that our implementation present better visualized result when compare to ImageVis3D. The dataset used in the experiment is CT images obtained from Osirix datasets for surgical repair of facial deformity.

Recent in 2015 apple present new graphic application programming interface Metal that can replace OpenGL ES. This study additionally project a light on the performance gained by Metal API over OpenGL ES API and how that can help in the bone visualization. Four dataset used to test the result, namely: Diaphysis, Distal epiphysis, Scapula and knee. The first three datasets used from the laboratory of human anatomy and embryology, University of Brussels (ULB), Belgium, the last one used from Osirix DICOM sample.