计算机应用研究2017，Vol.34Issue(11)：3291-3294,4.DOI:10.3969/j.issn.1001-3695.2017.11.020
基于反向投影的zero-shot learning目标分类算法研究
Zero-shot learning object recognition based on inverse projection
摘要
Abstract
Zero-shot learning(ZSL) aims to recognise new objects without having training samples of them.The kernel of the traditional regression method was to project the visual features into the semantic space,without taking full advantage of the sample information contained in the visual features,meanwhile the training computation was large.This paper proposed to project prototype to visual feature space,which was referred to as inverse projection.It had a very efficient closed-form solution.Extensive experiments on two benchmark datasets show that the proposed ZSL method significantly outperforms the state-of-the-arts.关键词
zero-shot learning/目标分类/反向投影/解析解Key words
zero-shot learning/objects recognition/inverse projection/closed-form solution分类
信息技术与安全科学引用本文复制引用
冯鹏,庹红娅,乔凌峰,王洁欣,敬忠良..基于反向投影的zero-shot learning目标分类算法研究[J].计算机应用研究,2017,34(11):3291-3294,4.基金项目
国家自然科学基金资助项目(61175028) （61175028）
