New spatial-organization-based scale and rotation invariant features for heterogeneous-content camera-based document image retrieval.

•We have extended our earlier proposed feature descriptor named Scale and Rotation Invariant Features (SRIF).•SRIF is capable of extracting and encoding discriminatory information from heterogeneous-content in document images.•SRIF is computed based on geometrical constraints between pairs of nearest points around a keypoint.•SRIF has built-in capability to robustly deal with feature point extraction errors.•SRIF outperforms the state-of-the-art in terms of processing time with an equal or greater Recall and Precision.