ERS transform for the detection of bronchi on CT of the lungs

Abstract

The identification of bronchi on Computed Tomography (CT) images of the lungs provides valuable clinical information for the assessment of patients with suspected bronchiectasis, emphysema, or constrictive obliterative bronchiolitis. The automated recognition of the airways is an important part of a diagnosis-aid system. It resolves potential ambiguities associated with intensitybased feature extractors. On CT images. cross-sections of bronchi normally appear as elliptical rings and this paper presents a novel technique for their recognition. The proposed method, the ERS transform. is based on the analysis of the distribution of edges in local polar co-ordinates. Pixels are ranked according to local edge (E) strength. radial (R) uniformity. and local symmetry (S). A discrete implementation of the technique is provided which reduce< the computational cost of the ERS transform by using a geometric approximation of the intensity patterns. The method compares favourably to other methods such as template matching or Hough transform. Noise-sensitivity of the technique was evaluated on a set of synthetic images and patient study was undertaken with a set of 27 cross-sectional images showing different lung pathologies. Agreement with an experienced radiologist was reached in 76 out of 136 bronchi (agreement rate: 57%). which suggests satisfactory statistical significance for using the ERS transform as part of a computerised diagnosis aid system.