Research On Semi-Automatic Image Annotation System Based On Segmentation And SVM

As high-resolution digital cameras become more affordable and widespread to the people, high-quality digital images become ever richer and more multiple. With the exponential growth on high-quality digital images, there is an urgent need to support more effective image retrieval over large-scale archives.However, content-based image retrieval (CBIR) is still immature and most existing CBIR systems can only support feature-based image retrieval. For most naive users, it is very hard for them to specify their query concepts by using the low-level visual features directly. And the existing image retrieval technique based on keyword fuses the virtue of textual retrieval, but it couldn't describe the semantic and character together to build the relation between keyword and image. Image retrieval aims searching known image information in large database, but the rapid increase of image and small collection of known image makes the database incomplete. Users prefer to get information from unknown image, so prediction of unknown image according to rule of existing data is more practical at present.Image annotation makes users explain their query with keyword easily and analyze the query from the known data. So the naive users can specify their query concepts easily by using the relevant keywords. The performance of image annotation largely depends on two inter-related issues: (1) suitable frameworks for image content representation and automatic feature extraction; (2) effective algorithms for image classifier training and feature subset selection.A novel approach to semi-automatically and progressively annotating images with keywords is presented in this paper. This plan by specifying the Using image segmentation and MPEG7 in describing features of landscape, this plan specifies the visual descriptors' scheme to solve the first issue. Then train features to category models using SVM (Support Vector Machine), and optimize models by relevant feedback, so the second issue is solved. The system implemented with this method has satisfied annotation results. The progressive annotation process is embedded in the course of integrated keyword-based and content-based image retrieval. It improves accuracy of annotation by user feedback. When users submit a query and receive keyword annotation. Then they provide relevance feedback. The annotated results which receive positive feedback are automatically added to the image set and facilitate keyword-based image retrieval in the future. The results with negative feedback are picked into correct categories to update system. The coverage and quality of image annotation in such a database system is improved progressively as the cycle of search and feedback increases.The strategy of semi-automatic image annotation is better than manual annotation in terms of efficiency and better than automatic annotation in terms of accuracy. A performance study is presented which shows that high annotation coverage can be achieved with this approach, and a preliminary user study is described showing that users view annotations as important and will likely use them in image retrieval. The user study also suggested user interface enhancements needed to support relevance feedback. We believe that similar approaches could also be applied to annotating and managing other forms of multimedia objects.