Contour and Texture Analysis for Image Segmentation
International Journal of Computer Vision
Using Pseudo-Relevance Feedback to Improve Image Retrieval Results
Advances in Multilingual and Multimodal Information Retrieval
Mining multi-tag association for image tagging
World Wide Web
Evaluation of image segmentation algorithms from the perspective of salient region detection
ACIVS'11 Proceedings of the 13th international conference on Advanced concepts for intelligent vision systems
Visual vocabulary optimization with spatial context for image annotation and classification
MMM'12 Proceedings of the 18th international conference on Advances in Multimedia Modeling
| Hi-index | 0.00 |
Owing the widespread use of digital image, methods of high efficiency of image retrieval from WWW are becoming urgent requirements to users. But the traditional search engines are mostly based on keywords. This paper presents a modular image search engine based on keywords and contents, which organically combines search engine technology of keywords and images' color feature. The system searches images from WWW by WEB robots, extracts their relevant contents and color features, and then stores them into a database. When a user gives a query, the system displays the results according to the user's search requirements. For the color features of an image, a quantified method based on the maximum pixels ratio of irregular connected regions is raised. Experiments show that the method improves the retrieval efficiency and can get an expected search result more accurately, so as to satisfy the customer's needs.