计算机辅助设计与图形学学报
計算機輔助設計與圖形學學報
계산궤보조설계여도형학학보
JOURNAL OF COMPUTER-AIDED DESIGN & COMPUTER GRAPHICS
2015年
2期
238-248
,共11页
史骏%姜志国%赵丹培%陆明
史駿%薑誌國%趙丹培%陸明
사준%강지국%조단배%륙명
图嵌入%低秩表示%半监督学习%维数缩减%人脸识别
圖嵌入%低秩錶示%半鑑督學習%維數縮減%人臉識彆
도감입%저질표시%반감독학습%유수축감%인검식별
graph embedding%low-rank representation%semi-supervised learning%dimensionality reduction%face recognition
针对基于图嵌入的鉴别投影方法对近邻参数的敏感以及实际应用中样本类别信息不足对图嵌入方法鉴别性能的影响,提出一种基于自适应近邻选择和低秩表示的半监督鉴别分析方法。该方法利用所有类内样本点构造类内图来描述类内样本的紧致性,借助最远类内样本的邻域自适应地选取该邻域内不同类样本点构造类间图,以描述类间样本的可分性;此外,利用低秩表示方法挖掘不带类别信息样本的潜在低秩结构,以保留样本的全局相似关系。在ORL和FERET人脸数据库上的实验结果,验证了文中方法的有效性及对噪声的鲁棒性。
針對基于圖嵌入的鑒彆投影方法對近鄰參數的敏感以及實際應用中樣本類彆信息不足對圖嵌入方法鑒彆性能的影響,提齣一種基于自適應近鄰選擇和低秩錶示的半鑑督鑒彆分析方法。該方法利用所有類內樣本點構造類內圖來描述類內樣本的緊緻性,藉助最遠類內樣本的鄰域自適應地選取該鄰域內不同類樣本點構造類間圖,以描述類間樣本的可分性;此外,利用低秩錶示方法挖掘不帶類彆信息樣本的潛在低秩結構,以保留樣本的全跼相似關繫。在ORL和FERET人臉數據庫上的實驗結果,驗證瞭文中方法的有效性及對譟聲的魯棒性。
침대기우도감입적감별투영방법대근린삼수적민감이급실제응용중양본유별신식불족대도감입방법감별성능적영향,제출일충기우자괄응근린선택화저질표시적반감독감별분석방법。해방법이용소유류내양본점구조류내도래묘술류내양본적긴치성,차조최원류내양본적린역자괄응지선취해린역내불동류양본점구조류간도,이묘술류간양본적가분성;차외,이용저질표시방법알굴불대유별신식양본적잠재저질결구,이보류양본적전국상사관계。재ORL화FERET인검수거고상적실험결과,험증료문중방법적유효성급대조성적로봉성。
Considering the discriminant projection methods based on graph embedding are sensitive to the neighbor parameter and the fact that there is no sufficient class-label information of samples in practical applica-tions which has an impact on the performance of graph embedding based methods, a semi-supervised discrimi-nant analysis method based on adaptive neighbor selection and low-rank representation is proposed. The method uses all the intraclass samples to construct the intraclass graph which can characterize the intraclass compactness, and simultaneously adaptively selects the interclass samples within the neighborhood produced by the farthest in-traclass sample to construct the interclass graph which is used to characterize the interclass separability. Further-more, the low-rank representation approach is applied to mine the latent low-rank structure of unlabeled samples and thus preserve the global similarity relationship of samples. Experimental results on ORL and FERET face da-tabases demonstrate the effectiveness of our method and the robustness to noise.