Computer-based estimation of circulating blood volume from ultrasound imagery

Karami, Ebrahim (2019) Computer-based estimation of circulating blood volume from ultrasound imagery. Doctoral (PhD) thesis, Memorial University of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (3MB)


Detection of relative changes in circulating blood volume is important to guide resuscitation and manage a variety of medical conditions including sepsis, trauma, dialysis and congestive heart failure. In recent years, ultrasound images of inferior vena cava (IVC) and internal jugular vein (IJV) have been used to assess volume status and guide fluid administration. This approach has limitations in that a skilled operator must perform repeated measurements over time. In this dissertation, we develop semi-automatic image processing algorithms for estimation and tracking of the IVC anterior-posterior (AP)-diameter and IJV crosssectional area in ultrasound videos. The proposed algorithms are based on active contours (ACs), where either the IVC AP-diameter or IJV CSA is estimated by minimization of an energy functional. More specifically, in chapter 2, we propose a novel energy functional based on the third centralized moment and show that it outperforms the functionals that are traditionally used with active contours (ACs). We combine the proposed functional with the polar contour representation and use it for segmentation of the IVC. In chapters 3 and 4, we propose active shape models based on ellipse; circle; and rectangles fitted inside the IVC as efficient, consistent and novel approaches to tracking and approximating the anterior-posterior (AP)-diameter even in the context of poor quality images. The proposed algorithms are based on a novel heuristic evolution functional that works very well with ultrasound images. In chapter 3, we show that the proposed active circle algorithm accurately, estimates the IVC AP-diameter. Although the estimated AP-diameter is very close to its actual value, the clinicians define the IVC AP-diameter as the largest vertical diameter of the IVC contour which deviates from its actual definition. To solve this problem and estimate the AP-diameter in the same way as its clinical definition, in chapter 4, we propose the active rectangle algorithm, where clinically measured AP-diameter is modeled as the height of a vertical thin rectangle. The results show that the AP-diameter estimated by the active rectangle algorithm is closer to its clinically measurement than the active circle and active ellipse algorithms. In chapter 5, we propose a novel adaptive polar active contour (Ad-PAC) algorithm for the segmentation and tracking of the IJV in ultrasound videos. In the proposed algorithm, the parameters of the Ad-PAC algorithm are adapted based on the results of segmentation in previous frames. The Ad-PAC algorithm has been applied to 65 ultrasound videos and shown to be a significant improvement over existing segmentation algorithms. So far, all proposed algorithms are semi-automatic as they need an operator to either locate the vessel in the first frame, or manually segment the first first and work automatically for the next frames. In chapter 6, we proposed a novel algorithm to automatically locate the vessel in ultrasound videos. The proposed algorithm is based on convolutional neural networks (CNNs) and is trained and applied for IJV videos. In this chapter we show that although the proposed algorithm is trained for data acquired from healthy subjects, it works efficiently for the data collected from coronary heart failure (CHF) patients without additional training. Finally, conclusions are drawn and possible extensions are discussed in chapter 7.

Item Type: Thesis (Doctoral (PhD))
Item ID: 14048
Additional Information: Includes bibliographical references.
Keywords: Ultrasound imaging, Image processing, Active contours, Internal jugular vein, Inferior Vena Cava, Deep learning
Department(s): Engineering and Applied Science, Faculty of
Date: October 2019
Date Type: Submission
Library of Congress Subject Heading: Blood volume--Measurement; Blood volume--Measurement--Mathematical models

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics