The Ultrasound Visualization Pipeline - Scientific Visualization: Uncertainty, Multifield, Biomedical, and Scalable Visualization - page 269

Data Type Year

Anatomy Technique

2010

Brain CT

Pre-processing

MRI

Segmentation

2005

Heart

Breast
Kidney

Doppler

Registration

2000

Liver

3D Ultrasound

Fetus
Pelvis

Rendering

1995

Freehand

Augmented
Reality

2D Ultrasound 1990

Fig. 24.2 The different classifications shown in a parallel-coordinate plot. The colors depict which
technique a publication has given the most weight

see an increase in rendering techniques for 3Dultrasound in the last five years. Volume
rendering is often considered to be a solved problem. However, our study shows that
much work dealt with volumetric ultrasound data. Yet, 3D ultrasound rendering can
still not be considered a solved problem. The high presence of noise, shadows from
hyper-echoic areas and inconsistent data values provide a great challenge to make
3D ultrasound a more easy-to-use tool for examiners.
We also see an absence of augmented reality techniques for 3D ultrasound. Yet
another trend is the neglection of 2D ultrasound from the visualization community.
2D ultrasound is the most used modality by physicians and while presenting the
signal data onto the screen is straight forward, understanding what you see is not
so trivial. Increasing the readability of 2D ultrasound is mostly worked on in the
commercial domain, aiming to give a company an edge over its rivals.
In Fig. 24.2 we see the categorized papers in a parallel coordinate plot where each
axis corresponds to the different taxonomy classification. The third axis (the pipeline
axis) is selected as the classification for this survey. Five categories where chosen
based on what we recognize as the essential parts in the visualization pipeline for
ultrasound data:

￿ Pre-processing : Processing ultrasound data prior to segmentation, registration or

rendering.

￿ Segmentation : Extracting features from ultrasound data.

￿ Registration : Combining ultrasound data with other types of medical imaging

modalities.

￿ Rendering : Presenting ultrasound data.

￿ Augmented Reality : Mixing ultrasound rendering with natural vision.

In the following sections we motivate the need for each of the major topics,
focusing on significant techniques and how they deal with the characteristics of
ultrasound data.