Fresh-sliced tissue inspection: Characterization of pork and salmon composition based on fractal analytics

Abstract

[EN] The capability of fractal analytics of digital images for characterizing sliced tissue was tested in fat and water fractions terms. The study was based on three factors: tissue type (pork loin and salmon); fractal parameter (fractal dimension and lacunarity); image type in color terms (grayscale and RGB channels). After capturing images of tissue samples, the water and fat fractions were analyzed from each one. The fractal information was expressed as specific spectra from each image type, which were related with composition properties. The relationship between both data blocks was tested by multivariate regression assays. The results reported that fractal information collected the variability due to both composition and tissue morphology. Both fractal parameters and all the image types successfully reported the results. The two fractal parameters displayed a strong dependency on components, and both characterized sample properties. No considerable differences were found in the results for image type. However, the inclusion of the entire color spectra increased the observed coefficients. The idea presented here could be a simple rapid non destructive technique for characterizing tissue composition.