Source: Radiomics | Radiology Reference Article | Radiopaedia.org
The process of extracting quantitative features from medical images (.dicom) can be briefly described as:
- First, some preprocessing steps are performed to increase the quality and usability of the image, e.g. contrast, edge enhancement etc.
- A region of interest (ROI) - area of interest (2D images) or volume of interest (3D images) - is identified either manually, semi-automated, or fully automated using artificial intelligence.
- High dimension features are extracted from these ROIs that include semantic and agnostic features.
- Semantic features are morphological features that are commonly used in radiology reports to describe lesions.
- Agnostic features are more complex mathematically extracted quantitative features.
- Examples of semantic features
- Equivalent examples of agnostic features
- kurtosis or skewness (of the image histogram)
- Haralick textures
- Laws textures
- Laplacian transforms
- Minkowski functions
- fractal dimensions
- These extracted features are then used to generate a report, which is placed in a database along with other data, such as clinical and genomic data such as genes, mutations, and expression patterns.
- Combining molecular and imaging metrics in cancer: radiogenomics | Insights into Imaging | Full Text (springeropen.com)
- Radiomics | Radiology Reference Article | Radiopaedia.org
Related: Magnetic resonance imaging (MRI)