Feasibility of Haralick's Texture Features for the Classification of Chromogenic In-situ Hybridization Images

摘要: This paper presents a proof of concept for the usefulness second-order texture features qualitative analysis and classification chromogenic in-situ hybridization whole slide images in high-throughput imaging experiments. The challenge is that currently, gold standard gene expression grading such expert assessment. idea research team to use different approaches these will be used structural segmentation functional expression. article perspective select number textural are going classification. In our experiment, natural grouping image samples (tiles) depending on their local properties was explored an unsupervised procedure. reduced two dimensions with fuzzy c-means clustering. overall conclusion this experiment Haralick viable choice insitu data. principal component approach produced slightly more “understandable” from annotator’s point view classes.