A Study On Applying Content-based Image Retrieval Technology To The Management Of DLRs

Information technology, which is based on multimedia and network technology, is very important for social development and becomes a tool of enhancing human creative ability. With the rapid development of information technology, which is widely used in the field of education, education technology has been reformed in its all aspects. As a result, the concept of the digital learning is received more concerns by relevant researchers. Digital learning is a new way of learning under the condition of the rich resources provided by the information technology and the new communication methods. The development and application of digital learning resources plays an important role in digital learning.Digital Learning Resources (DLR) includes the text, image, video and other multimedia information resources which are processed by various digital techniques. Users can employ these resources with information technology. The characters of DLR are multimedia, hypertext, friendly interaction, virtual simulation and remote share. And the relationship of knowledge is a network. DLR is now welcomed by most learners and widely used in the area of education. However, while various resources are increased surprisedly, there are many problems in the use and management of these huge amounts of resources.The traditional methods of image retrieval mainly use text-based image retrieval technology, which is developed maturely. But the image index information needs to be annotated manually, which is a time-consuming work. Furthermore, there is a problem that the image annotations may not be precisely described. These factors may affect the management and use of resources. Therefore, after the analysis of the circumstances of DLR management, we propose a content-based image retrieval algorithm to solve the existing problems in text-based image retrieval technology. Sine video is consisted of images, the video indexing is also researched to some extent in the dissertation in order to enrich the content of DLR management.The innovations in this dissertation are to combine multiple features of images such as skin texture, Gaussian density for image retrieval, and to employ shot boundary detection, key frame extraction and multi-feature image retrieval for video indexing. We apply the content-based image retrieval and video indexing algorithms into the DLR management in order to solve traditional image and video retrieval problems.In the validation experiments, we established the image database consisting of the JPEG images and video database consisting of MPEG video. The experimental results show that the proposed algorithm outperforms in terms of recall and precision compared with exiting techniques. After finished experiments and summarized the work of the dissertation, we provide some suggestions for DLR management in future work.