Abstract: In this paper, we study the automatic classification and identification of paper cutting images to find a new way to realize the efficient classification, storage and retrieval of paper cutting images. The texture, forms and combined features of paper cutting images were extracted to make simulation experiment of identifying the paper cutting partitioned according to different cultural connotations. The results showed that classification of texture and form features for paper cutting images extracted by means of gray cooccurrence matrix and form invariable matrix can lead to good classification effect in the classification experiment.

Key words: Paper cutting image, Feature extraction, SVM classifier, Image recognition