TY - JOUR
T1 - Interactive object-based image retrieval and annotation on iPad
AU - Han, Junwei
AU - Xu, Ming
AU - Li, Xin
AU - Guo, Lei
AU - Liu, Tianming
PY - 2014/10
Y1 - 2014/10
N2 - Apple iPad is a portable tablet computer that offers users a generic platform for consumer media including games, books, and movies. Though iPad is gaining popularity very quickly, its application in content-based image retrieval and annotation is still in its infancy. This paper aims to develop an interactive system to efficiently retrieve and annotate image objects on iPad, which mainly consists of two components of the front-end GUI (graphical user interface) and the back-end retrieval model. In the first component, an iPad-based GUI is implemented, which can provide users with an efficient way to select query objects and facilitate annotations. In the second component, we propose an object-based image retrieval algorithm that combines a novel feature descriptor based on context-preserving bags-of-words (BoW) and a two-stage re-ranking technique to measure the similarity between the query image and each image in the database. The retrieval results are returned and visualized on the iPad-based GUI, and annotations offered by users can be propagated among them. The communication between the front-end GUI and the back-end module is through the use of wireless networks. Comprehensive experiments on several benchmark datasets demonstrated the effectiveness of the proposed framework.
AB - Apple iPad is a portable tablet computer that offers users a generic platform for consumer media including games, books, and movies. Though iPad is gaining popularity very quickly, its application in content-based image retrieval and annotation is still in its infancy. This paper aims to develop an interactive system to efficiently retrieve and annotate image objects on iPad, which mainly consists of two components of the front-end GUI (graphical user interface) and the back-end retrieval model. In the first component, an iPad-based GUI is implemented, which can provide users with an efficient way to select query objects and facilitate annotations. In the second component, we propose an object-based image retrieval algorithm that combines a novel feature descriptor based on context-preserving bags-of-words (BoW) and a two-stage re-ranking technique to measure the similarity between the query image and each image in the database. The retrieval results are returned and visualized on the iPad-based GUI, and annotations offered by users can be propagated among them. The communication between the front-end GUI and the back-end module is through the use of wireless networks. Comprehensive experiments on several benchmark datasets demonstrated the effectiveness of the proposed framework.
KW - iPad
KW - Multi-Touch
KW - Object Retrieval and Annotation
KW - Spatial Context
KW - Visual Bag-of-Words
UR - http://www.scopus.com/inward/record.url?scp=84878495900&partnerID=8YFLogxK
U2 - 10.1007/s11042-013-1509-6
DO - 10.1007/s11042-013-1509-6
M3 - 文章
AN - SCOPUS:84878495900
SN - 1380-7501
VL - 72
SP - 2275
EP - 2297
JO - Multimedia Tools and Applications
JF - Multimedia Tools and Applications
IS - 3
ER -