|
| |
Imaging and Tracking.
Introduction
One of the main objectives
within MAESTRO project is the development of imaging software tools to improve
the radiotherapy treatment planning and devices.
This includes image processing algorithms for automatic patient positioning and
monitoring the internal organ motion and region of interest (ROI) during IMRT
treatment. To this end methods based on image segmentation, active contour
models for ROI tracking and modelling together with filtering techniques for
movement prediction, registration and 3D reconstruction techniques are being
considered.
|
Organ Motion Prediction |
|
The tracking system is formed for two
subsystems: a segmentation system and an estimation one. |
|
In order to reach “Real Time”
applications the segmentation system must be a high speed segmentation
algorithm. Thus, the estimation system is based on Kalman filter with
four and/or six states. |
The results in off
line applications are quite promising in terms of computational time
and accuracy.
|
![]() |
|
Video 1. 8 photogram prediction
superimposed onto the original image sequence Video duration
10,47 s – Computing time 5.99 s |
|
|
![]()
|
|
Video 2. 6 photogram prediction
superimposed onto the original image sequence Video duration
8.80 s – Computing time 5.94 s
|
|
Segmentation and Tracking |
|
The
segmentation and tracking are based on modelling techniques founded on
active contours models. The method has a good performance on 3D CT set
of the thoracic and pelvic area. |
|
|
|
3D
Reconstruction |
|
3D reconstruction and rendering have
been achieved with the ROI segmented for the pelvic and thoracic area.
That is, the rectum, bladder, lung and spinal cord. |
|
|
|
Pelvic
Area
|
Thoracic
area
|
|
|
|
 |
|
|
Spinal cord
|
Lung
|
|
Acknowledgements
The group UCLM-ISA
(http://gvc.ind-cr.uclm.es/) is
thankful to GRUPO IMO for their help and support.
 |
