Indoor Scene Classification Based On Image Segmentation And Sparse Representation-based Classifier

Scene classification is a new topic in the field of image processing research since 90 th.Indoor scene classification is a research domain of scene classification, which is also a research challenge of scene classification. Once the accuracy of indoor scene classification is improved effectively, it is bound to play a more important role in the field of image retrieval, intelligent robots and video retrieval. Therefore, the study of indoor scene classification is very important and meaningful.In this paper, a new system based on the sparse representation classification was proposed, after studying the theoretical knowledge of pattern recognition theory. This system is based on the theory of image segmentation and sparse representation.The results show that our system can improve the classification accuracy to a certain extent. The main content and key technology on our research work is as followings.First of all, this paper reviews the background knowledge and the significance of the research of indoor scene classification and sparse representation and introduces the current research status of the indoor scene image, and the sparse representation classification method was applied to the indoor scene image classification. Secondly, to further improve the indoor scene image classification technology based on the theory of sparse representation, the idea of image segmentation was applied to this classification system. Study of Otsu threshold segmentation method, and the principle of the original Otsu algorithm and proposes an improved method of Otsu algorithm.Comparative experiment was designed with other classification,the results and data was analyzed shows that our system can improve the classification accuracy to a certain extent.