We deal with a general task of an image retrieval, more
specifically, we focus on searching for an input image
in a large database of images. The original motivation
came from an image plagiarism detection. We propose a
searching algorithm based on the technique of F-transform.
Our main requirements for the algorithm are computational
speed and success rate. The F-transform is used in
preprocessing to reduce the domain dimension and by this
it significantly speeds up the whole process. We present
several experiments and measurements which demonstrate the
power and effectiveness of the proposed method.