8  Pyradiomics Overview

8.1 PKG Structures

8.1.1 Class Hierarchy of pyradiomics

  1. radiomics.featureextractor.RadiomicsFeatureExtractor

    • Description: The central class used for feature extraction. This class handles image loading, pre-processing, feature calculation, and results aggregation.
    • Key Methods:
      • execute(): Extracts features from an image and a mask.
      • enableFeatures(): Enables specific feature classes for extraction.
      • disableFeatures(): Disables feature extraction.

  2. radiomics.featurebase.RadiomicsFeaturesBase

    • Description: The base class for all feature classes in Pyradiomics. All feature classes inherit from this base class.
    • Child Classes:
      • radiomics.firstorder.RadiomicsFirstOrder
        • Extracts first-order statistics such as mean, variance, skewness, etc.
      • radiomics.glcm.RadiomicsGLCM
        • Extracts Gray Level Co-occurrence Matrix (GLCM) features (texture features).
      • radiomics.gldm.RadiomicsGLDM
        • Extracts Gray Level Dependence Matrix (GLDM) features.
      • radiomics.ngtdm.RadiomicsNGTDM
        • Extracts Neighborhood Gray Tone Difference Matrix (NGTDM) features.
      • radiomics.glszm.RadiomicsGLSZM
        • Extracts Gray Level Size Zone Matrix (GLSZM) features.
      • radiomics.glrlm.RadiomicsGLRLM
        • Extracts Gray Level Run Length Matrix (GLRLM) features.
      • radiomics.shape.RadiomicsShape
        • Extracts shape-based features (e.g., volume, surface area, etc.).

  3. radiomics.imageoperations.ImageOperations

    • Description: This class handles image preprocessing operations like resampling, filtering, and normalization.
    • Key Methods:
      • resampleImage(): Resamples the image based on a given spacing.
      • applyFilter(): Applies filters like wavelets or Laplacian of Gaussian (LoG) to the image.

  4. radiomics.generalinfo.GeneralInfo

    • Description: Class responsible for extracting general information about the image (e.g., image dimensions, voxel size, and acquisition details).

  5. radiomics.logger.RadiomicsLogger

    • Description: Handles logging of the pyradiomics execution, including errors and progress reporting.
    • Key Methods:
      • setLevel(): Sets the logging level (e.g., DEBUG, INFO, WARNING).
      • addHandler(): Adds a logging handler to direct logs to a file or console.

  6. radiomics.calculate.CalcFeatures

    • Description: Contains the lower-level logic for calculating features. It includes utility functions to perform calculations within the different feature classes.
    • Key Methods:
      • calculateFeatures(): Central method to perform feature calculation for a given feature class.

8.1.2 High-Level Overview of the Class Structure

RadiomicsFeatureExtractor
│
├── RadiomicsFeaturesBase
│   ├── RadiomicsFirstOrder
│   ├── RadiomicsGLCM
│   ├── RadiomicsGLDM
│   ├── RadiomicsNGTDM
│   ├── RadiomicsGLSZM
│   ├── RadiomicsGLRLM
│   └── RadiomicsShape
│
├── ImageOperations
│   ├── resampleImage()
│   └── applyFilter()
│
├── GeneralInfo
│   └── extractGeneralInfo()
│
└── RadiomicsLogger
    ├── setLevel()
    └── addHandler()

8.1.3 Summary of Key Classes

  • RadiomicsFeatureExtractor: The main interface for users to extract features.
  • RadiomicsFeaturesBase: Parent class for all feature extraction classes.
  • Specific Feature Classes (e.g., GLCM, First Order): Responsible for extracting different sets of radiomics features.
  • ImageOperations: Handles image preprocessing steps (e.g., filtering and resampling).
  • GeneralInfo: Extracts general metadata from the image.
  • RadiomicsLogger: Manages logging for the package.