Changes of gray level co-occurrence matrix (GLCM) texture and antioxidant level during freezing of pumpkin

Abstract

GLCM (gray level co-occurrence matrix) textural features are rarely extracted from SEM (scanning electron microscopy) food images to study the effect of food processing on cellular microstructure and bioactive contents. Using the GLCM approach, the current study attempted to use textural information from pumpkin SEM images to obtain detailed cellular degradation caused by freezing treatment and to find their relationship with the antioxidant content. High dimensional textural features of the samples were obtained during image processing and subjected to multivariate analysis. The increase of antioxidant activity due to the freezing of pumpkin was used to create a vector factor to indicate the membership class of individual pumpkin samples. Following freezing, the flavonoid content of the pumpkins increased ranging from 15.27% to 70.39%. Freezing also caused the formation and increased of several volatile compounds detected by GCMS. The PCA (principal component analysis) results showed that two components were sufficient to explain the most variance of the dataset. Further supervised sPLS-DA (sparse partial least squares-discriminant analysis) indicated a clear separation of samples based on their antioxidant levels, suggesting an existing relationship between the texture and antioxidant. In conclusion, GLCM textural changes can be used accurately to examine the effect of freezing on the antioxidant of pumpkin.