自适应学习的多特征元素协同表示分类算法

Self-adaptive learning algorithm for collaborative representation classification of multi-feature elements

针对基于传统的稀疏表示分类算法的单特征鉴别性较弱这一不足,提出一种基于自适应学习的多特征元素协同表示分类算法SLMCE_CRC。该算法结合多特征子字典的思想,对样本提出特征元素的双重分解,并分别从特征和元素角度分别进行相应的协同表示,自适应地学习出各个特征的稀疏权重和元素的残差权重,并进行线性加权,从而实现目标的分类。实验结果表明,使用该方法能显著提高识别率,尤其对含有较多特征细节的图像数据,具有一定的实用价值。

To address the weak discriminative power of Sparse Representation Classification （SRC）,a self-adaptive learning algorithm for collaborative representation classification of multi-feature elements named SLMCE_CRC was proposed.Based on the idea of multi-feature sub-dictionary,the sample was collaboratively represented by features and elements,the sparse weights of features and the residual weights of elements were learnd self-adaptively and combined linearly to classify the samples.The experimental results demonstrate the effectiveness and high classification accuracy of the proposed algorithm.It is suitable to images with multi-features.