BriefGPT.xyz
Aug, 2012
信息论字典学习的图像分类
Information-theoretic Dictionary Learning for Image Classification
HTML
PDF
Qiang Qiu, Vishal M. Patel, Rama Chellappa
TL;DR
本研究提出一种两阶段方法，基于信息最大化原理学习对象分类任务的字典。第一阶段选择单词符号以最大化相互信息度量字典的紧凑性、辨别性和重构性能；第二阶段使用简单的梯度上升算法更新所选择的单词符号以提高其重构能力和辨别能力。实验结果表明该方法在实现基于图像的分类任务方面有效。
Abstract
We present a two-stage approach for learning dictionaries for
object classification
tasks based on the principle of
information maximization
. The proposed method seeks a dictionary that is compact, discriminative
→