摘要
基于协同表示的分类(CRC)以其卓越的协同能力成为人脸分类领域的一个突破。然而在实际应用中,通常只提供很少甚至是单个人脸图像来进行人脸识别,这导致了CRC无法很好地处理光照、表情、姿态和遮挡等问题。针对该问题,提出一种判别性双向协同表示的图像识别算法(DB-CRC)。首先通过引入判别式字典学习(FDDL)模型学习得到一个结构化字典,使得每个特定类的子字典对相关类的样本具有良好的表示能力,由此,较大的类间离散度和较小的类内离散度使得重构误差和编码系数都具有判别性;然后将学习得到的稀疏编码系数作为测试样本数据进行双向表达,建立快速逆向表示模型,利用双向表示策略估计每个测试样本与结构化字典之间的双向重构残差信息;最后利用竞争融合方法对来自双向表示模型的重构残差进行加权排名,实现最终的人脸分类。在AR、PIE、LFW等通用人脸数据库上的实验结果验证了该算法的有效性,特别是对小样本问题的鲁棒性。
Collaborative representation-based classification(CRC)has become a breakthrough in the field of face classification due to its effective collaborative capability.However,for some practical scenarios,it used few or even a single face image to address the issues such as illumination,expression,pose and occlusion changes,which might lead to inadequate capability for CRC.To solve this problem,this paper proposed a discriminative bi-directional CRC(DB-CRC)algorithm for image recognition.Firstly,it obtained a structured dictionary using Fisher discrimination dictionary learning(FDDL)model,in which the sub-dictionary of specific-class had robust representation capability for the samples of related-classes.Therefore,by means of larger inter-class and smaller intra-class scatters,the reconstruction error and coding coefficient could be described in discriminant manner.Then,it used the obtained sparse coding coefficients as the test samples for the bi-directional representation,and established a fast inverse representation model.It estimated the bi-directional reconstruction based residual information between each test sample and the structured dictionary via bi-directional representation strategy.Finally,this paper used the competitive fusion method to achieve the final classification result,by weighting and ranking the obtained reconstructed resi-duals from the bi-directional representation model.Experimental results conducted on a set of well-known face databases including AR,PIE and LFW verify the effectiveness of this algorithm,especially its robustness for small sample size problem.
作者
王亚楠
宋晓宁
Wang Yanan;Song Xiaoning(School of Artificial Intelligence&Computer,Jiangnan University,Wuxi Jiangsu 214122,China)
出处
《计算机应用研究》
CSCD
北大核心
2021年第2期615-618,共4页
Application Research of Computers
基金
国家重点研发计划资助项目(2017YFC1601800)
国家自然科学基金资助项目(61876072)
中国博士后科学基金特别资助项目(2018T110441)
江苏省六大人才高峰项目(XYDXX-012)。
