An Automatic Wavelet-Based Approach for Lung Segmentation and Density Analysis in Dynamic CT
- Resource Type
- Conference
- Authors
- Talakoub, Omid; Helm, Emma; Alirezaie, Javad; Babyn, Paul; Kavanagh, Brian; Grasso, Francesco; Engelberts, Doreen
- Source
- 2007 IEEE Symposium on Computational Intelligence in Image and Signal Processing Computational Intelligence in Image and Signal Processing, 2007. CIISP 2007. IEEE Symposium on. :369-374 Apr, 2007
- Subject
- Signal Processing and Analysis
Computing and Processing
Wavelet analysis
Lungs
Computed tomography
Image edge detection
Diseases
Ventilation
Motion pictures
Rabbits
Image analysis
Image segmentation
- Language
Acute respiratory distress syndrome (ARDS) can occur in people with or without previous lung disease. Analysis of aeration in artificial ventilation for ARDS is one of the major applications of Computed Tomography (CT) lung density examination. A movie of an affected rabbit lung over the respiratory cycle was produced by dynamic CT with a cine loop technique. This technique can produce thousands of CT images for analysis with a single experiment. A fully automated algorithm based on the capability of wavelet transformation to detect edges in the image is proposed. This method accurately and consistently segments the lung in pulmonary CT images. The speed and accuracy of this technique allows it to outperform other methods when dealing with the large number of images created by dynamic Computed Tomography.