Title :
A Database Schema for Large Scale Annotated Image Dataset
Author :
Peng, Shaowu ; Liu, Leyuan ; Yang, Xiong ; Sang, Nong
Abstract :
This paper gives a detailed description of the database construction for LHI image dataset. LHI image dataset is an integration of many desired annotated tasks in an all-in-all labeling frame work, such as segmentation, sketch, and hierarchical decomposition, employing the And-Or graph grammar. A two-tier database is designed to manage large amount of various data and And-Or graph model, togethere with the complex relations between data and models. With the help of And-Or graph and two-tier database, precise, flexible and various types of data query and export are achieved, which can not be provided by other datasets.
Keywords :
Artificial intelligence; Computer vision; Humans; Image databases; Labeling; Large-scale systems; Pattern recognition; Signal processing; Signal processing algorithms; Visual databases; And-Or Graph; Image Annotation; Image Database; Image Dataset; Image Parsing;
Conference_Titel :
Image and Signal Processing, 2008. CISP '08. Congress on
Conference_Location :
Sanya, China
Print_ISBN :
978-0-7695-3119-9
DOI :
10.1109/CISP.2008.49