Conference Paper

Automated 3D Segmentation of Lung Fields in Thin Slice CT Exploiting Wavelet Preprocessing.

DOI: 10.1007/978-3-540-74272-2_30 Conference: Computer Analysis of Images and Patterns, 12th International Conference, CAIP 2007, Vienna, Austria, August 27-29, 2007, Proceedings
Source: DBLP


Lung segmentation is a necessary first step to computer analysis in lung CT. It is crucial to develop automated segmentation
algorithms capable of dealing with the amount of data produced in thin slice multidetector CT and also to produce accurate
border delineation in cases of high density pathologies affecting the lung border. In this study an automated method for lung
segmentation of thin slice CT data is proposed. The method exploits the advantage of a wavelet preprocessing step in combination
with the minimum error thresholding technique applied on volume histogram. Performance averaged over left and right lung volumes
is in terms of: lung volume overlap 0.983 ± 0.008, mean distance 0.770 ± 0.251 mm, rms distance 0.520 ± 0.008 mm and maximum
distance differentiation 3.327 ± 1.637 mm. Results demonstrate an accurate method that could be used as a first step in computer
lung analysis in CT.

5 Reads
  • Source
    • "Unfortunately, an objective comparison of their performance is not conceivable, or at least, not fair, due to the lack of an accepted manual segmentation common dataset. Nevertheless, some approaches are well-established in the literature in order to measure the results from automatic and manual processes, such as the relative intersection between the areas (Korfiatis et al. 2007), the Hausdorf distance (Ma et al. 2011) or the receiver operating characteristic (ROC) analysis (Gruszauskas et al. 2009, Gruszauskas et al. 2008). "
    [Show abstract] [Hide abstract]
    ABSTRACT: In recent years, the segmentation, i.e. the identification, of ear structures in video-otoscopy, computerised tomography (CT) and magnetic resonance (MR) image data, has gained significant importance in the medical imaging area, particularly those in CT and MR imaging. Segmentation is the fundamental step of any automated technique for supporting the medical diagnosis and, in particular, in biomechanics studies, for building realistic geometric models of ear structures. In this paper, a review of the algorithms used in ear segmentation is presented. The review includes an introduction to the usually biomechanical modelling approaches and also to the common imaging modalities. Afterwards, several segmentation algorithms for ear image data are described, and their specificities and difficulties as well as their advantages and disadvantages are identified and analysed using experimental examples. Finally, the conclusions are presented as well as a discussion about possible trends for future research concerning the ear segmentation.
    Computer Methods in Biomechanics and Biomedical Engineering 09/2012; 17(8). DOI:10.1080/10255842.2012.723700 · 1.77 Impact Factor
  • Source
    • "Segmentation of the lung volumes is a required preliminary step to lung tissue categorization [53]. The result of this step is a binary mask M lung that indicates the regions to be analyzed by the texture analysis routines. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We propose near-affine-invariant texture descriptors derived from isotropic wavelet frames for the characterization of lung tissue patterns in high-resolution computed tomography (HRCT) imaging. Affine invariance is desirable to enable learning of nondeterministic textures without a priori localizations, orientations, or sizes. When combined with complementary gray-level histograms, the proposed method allows a global classification accuracy of 76.9% with balanced precision among five classes of lung tissue using a leave-one-patient-out cross validation, in accordance with clinical practice.
    IEEE transactions on information technology in biomedicine: a publication of the IEEE Engineering in Medicine and Biology Society 05/2012; 16(4):665-75. DOI:10.1109/TITB.2012.2198829 · 2.49 Impact Factor
  • Source
    • "Als Beispiel wird an dieser Stelle auf die Ausführungen von Tönnies [14] verwiesen, der in seinem Buch eine detaillierten Überblick über grundlegende Segmentierungs-Verfahren gibt. Verschiedene 3D-Segmentierungsverfahren aus speziellen Fotos (wie MRT-oder CT- Bilder) wurden bereits in [6], [13], [3] und [2] vorgestellt. Da es sich bei diesen Arbeiten jedoch stets um kalibrierte Bildserien handelt, eignen sich diese Verfahren nicht für beliebige Kameras/Fotos und somit nicht für unser Verfahren. "
    [Show abstract] [Hide abstract]
    ABSTRACT: a) Beliebige Fotos (b) Bilderwelt (Vogelperspekti-ve) (c) Bilderwelt mit eingeblende-ten 3D-Rekonstruktionspunkten (d) 3D-Voronoi-Diagramm mit 3D-Informationen für jedes Seg-ment (e) Schematischer Aufbau des 3D-Segmentbildes (f) 3D-Segmentbild (Vogelpers-pektive) Abbildung 1: Erstellung eines 3D-Segmentbildes in einer Bilderwelt. Kurzfassung Dieser Beitrag stellt eine neuartige 3D-Segmentierung von Fotos durch extrahierte 3D-Informationen eines Structure-From-Motion-Algo-rithmus vor. Mithilfe der 3D-Informationen wird eine 3D-Bilderwelt erzeugt für die Darstellung der Fotos sowie der 3D-Rekonstruktionspunkte im 3D-Raum (Abb. 1(b) und Abb. 1(c)). Als Grundlage dienen beliebige, unka-librierte Fotos (Abb. 1(a)). Das Ziel dieser Arbeit ist die Zerlegung eines Fotos in logisch zusammenhängende Bereiche in Abhängigkeit der extrahier-ten 3D-Informationen der abgelichteten Szene (Abb. 1(d)). Die entstehenden 3D-Segmente werden an die entsprechende 3D-Position im Raum verschoben (Abb. 1(e)). Es entsteht ein 3D-Segmentbild (Abb. 1(f)).
Show more

Similar Publications