Answers are grounded in this paper and cite their sources.
This study's objective is to execute successful segmentation of the lung anatomy of HRCT of patients who have ILD and evaluate the segmentation performance. Initial segmentation process involved Otsu grey level thresholding and morphological filtering. Some of the problems encountered were the appearance of connected lungs because the left lung and right lung were very close to each other, and heavily diseased lungs with too much damaged tissue. By using Radon transform and accumulating the pixel width, a separation region could be found to split left lung and right lung. The separation process yielded 70.1% improvement for all the samples with connected lungs. This causes the segmentation results for Level 1 to increase to 79.01% for Right Lung and 92.59% for Left Lung, for level 2, successful segmentation increased to 88.89% for both lungs. In Level 3 segmentation results increased to 85.19% for Right Lung and 82.72% for Left Lung. For Level 4 segmentation results increased to 96.30% for Right Lung and 95.06% for Left Lung. For Level 5 segmentation results increased to 92.59% for Right Lung and 86.42% for Left Lung. Samples that could not be separated were due to the size of the splitting region which could be tackled with an adaptive splitting region in future works.
No takes yet. Share an insight, caveat, or question.
Than et al. (2014) studied this question.
Synapse has enriched 9 related papers. Consider them for comparative context: