Skip to Main Content

Analyzing Gray Level Co-occurrence Matrix texture features for Classifying Images

Analyzing Gray Level Co-occurrence Matrix texture features for Classifying Images
Analyzing Gray Level Co-occurrence Matrix texture features for Classifying Images

Category: Research Poster

Author(s): Mussa Hassen

Presenter(s): Mussa Hassen

Mentors(s): Emily King

Gray Level Co-occurrence Matrix (GLCM) texture features, or Haralick texture features, are used to quantify the texture of an image. A GLCM takes a gray scale image and looks at how often pairs of pixel values appear next to each other. It counts these pairs for a chosen distance and direction to create a matrix of all pixel pair combinations. These matrices are then aggregated and normalized. This normalized co-occurrence matrix is then used to compute several statistical texture features that summarize the image’s texture. This preliminary work introduces the varieties of GLCM constructions; some characteristics of these features; and interpretations of the features through their corresponding optimal GLCM. Then it finalizes with a useful application of this mathematical tool as inputs for an interpretable machine learning model.