Date on Master's Thesis/Doctoral Dissertation

12-2020

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Electrical and Computer Engineering

Degree Program

Electrical Engineering, PhD

Committee Chair

Farag, Aly

Committee Co-Chair (if applicable)

Ali, Asem

Committee Member

Alphenaar, Bruce

Committee Member

Li, Hongxiang

Committee Member

Zhang, Hui

Committee Member

Dryden, Gerald

Committee Member

Seow, Albert

Author's Keywords

visualization; tubular objects; virtual colonoscopy; segmentation

Abstract

In this dissertation, visualization for tubular objects, i.e., projecting 2D images from 3D inner surfaces of tubular objects, is investigated. Given surface points on 3D objects, an approach that most accurately and effectively projects 2D images from the 3D surface with minimal loss of information is desired. A new visualization method for tubular surfaces is proposed, denoted by "Fly-In". The approach uses a virtual camera ring that moves along the inner surface's centerline, obtaining projections of the surrounding views, forming small 3D topological rings within the tube rendered as a 2D rectangular image. A new visualization loss measure is also presented, which utilizes the camera axis's projection direction, the surface normal, and the camera focal length and surface distance ratio. This measure provides a more realistic interpretation of the visible surface in the camera's visualization frustum. A comparative study of Fly-in versus two state-of-the-art methods (Fly-Through and Fly-Over) is conducted on synthetic tubular objects and 3D reconstructed colons from CT Colonography (CTC) patient data. Quantitative results are computed using the new loss measure and visualized using color-coding. Theoretical and experimental results show the high performance of the proposed Fly-In compared to the state-of-art. We also propose a transformation method that embeds the 3D colon centerline into a 2D plane or a 1D line to solve the camera movement's stabilization issues inside the colon. We can use this transformation to stabilize the camera movement in different VC visualization methods (i.e., not only Fly-In). We use the visualization loss measure to quantify the proposed method's feasibility by comparing the visualization quality before and after applying the proposed transformation. Finally, we created an automatic segmentation method for colon surface from CT 3D images to provide a full pipeline for virtual colonoscopy. Experts in the CTC field evaluated our visualization system, and we show the system's efficiency based on the experts' opinion and their visualization results.

Download

Included in

Biomedical Commons

Share

COinS