Texture analysis in gel electrophoresis images using an integrative kernel-based approach

How to cite this article: Fernandez-Lozano, C. et al. Texture analysis in gel electrophoresis images using an integrative kernel-based approach. Sci. Rep. 5, 19256; doi: 10.1038/srep19256 (2015). <br>In order to generate the dataset, ten 2-DE images of different types of tissues and different experimental conditions were used. These images are from the dataset owned by G.-Z Yang (Imperial College of Science, Technology and Medicine, London) and have been used in several publications. For each image, two different clinicians agreed on 100 regions of interest manually segmented that were selected to build a training set with 1000 samples and 274 textural variables. We considered six groups of textural features: Histogram-based (first-order statistical texture features), Absolute Gradient, Run-length Matrix (high-order statistical texture features), Co-occurrence Matrix (second-order statistical texture features), Autoregressive Model and Wavelet. These features are based on image histogram, co-occurrence matrix (information about the grey level value distribution of pairs of pixels), image gradients (spatial distribution of grey level values), auto-regressive models (description of texture based on statistical correlation between pairs of pixels) and wavelet analysis (information about image frequency at different scales). We calculated those features with a specialized software called Mazda. Various approaches have demonstrated the effectiveness of this software, extracting textural features in different types of medical images