Local and global transformations to improve learning of medical images applied to Chest Radiographs

摘要

A potential drawback of computer-aided diagnosis (CAD) systems is that they tend to capture the noise characteristicsalong with signal variations due to a limited number of sources used in training. This leads to adecrease in performance on data from different sources. The variations in scanner settings, device manufacturersand sites pose a significant challenge to the learning capabilities of the CAD systems like chest radiographs, alsocalled Chest X-rays (CXR). In the proposed work, we investigate if preprocessing transformations like globalnormalization along with local enhancements are good to tackle the variability of data from multiple sourceson a supervised CXR classification system. We also propose a detail enhancement filter to enhance both finerstructures and opacities in CXRs. With the proposed preprocessing improvement, experiments were performedon 13,000 images across 3 public and one private data source using Dense Convolutional Network (DenseNet).The sensitivity at equal error rate (mean - sd) improved from 0:888 - 0:043 to 0:931 - 0:030 by applying acombination of global histogram equalization with the proposed detail enhancement filter when compared to theraw images. We conclude that the proposed transformations are effective in improving the learning of CXRsfrom different data sources.