Statistical image analysis and escort histograms in characterization of articular cartilage repair in a skeleton animal model

Statistical image analysis of an ensemble of digital images of histological samples is performed as an auxiliary investigation a result of the recently proposed method of articular cartilage repair utilizing growth plate chondrocytes in a skeleton animal model. A fixed–shift model of maximal likelihood estimates of image histograms applied for monochromatic (grayscale) images or their RGB components confirms the statistically significant effect of the previously proposed medical treatment. The type of staining used to prepare images of histological samples is related to the visibility of the effectiveness of medical treatment. Hellinger distance of escort distributions for maximal likelihood estimates of image histograms of medically treated and control samples is investigated to identify grayscale (or RGB) intensities responsible for statistically significant difference of the estimates. A difference of Shannon entropy quantifying informational content of the histograms allows one to identify staining and image colors which are most suitable to visualize cluster formation typical for articular cartilage repair processes.

Statistical image analysis for information system

The surface Kirchhoff integral is numerically calculated by the fourth-order Newton – Cotes method with high accuracy. The intensity is normalized. Implemented quantization of the signal with a unit step for different distances of observation is carried out. The cross section of the central topological object intensity and its characteristics are shown. Variation curves and cumulates are constructed for intensity distributions at different observation distances. The basic statistical parameters of distributions are determined, such as, the mode, the mean, the median and the standard deviation.