HEp-2 cell classification with Vector of Hierarchically Aggregated Residuals

详细信息    查看全文

• 作者：Dimitris Kastaniotis^a ; ^{dkastaniotis@upatras.gr}Author Vitae ; Foteini Fotopoulou^b ; ^{fotopoulou@ceid.upatras.gr}Author Vitae ; Ilias Theodorakopoulos^a ; ^{iltheodorako@upatras.gr}Author Vitae ; George Economou^a ; ^{economou@physics.upatras.gr}Author Vitae ; Spiros Fotopoulos^a ; ^{spiros@physics.upatras.gr}Author Vitae
• 关键词：HEp-2 cell classification ; Feature encoding ; Feature aggregation ; Sparse representation
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：May 2017
• 年：2017
• 卷：65
• 期：Complete
• 页码：47-57
• 全文大小：685 K
• 卷排序：65

文摘

A novel method for encoding features’ Sparse Representation residuals is proposed. Multiple levels of unsupervised learning are utilized. Local gradient descriptors are encoded into fixed length vectors. The method is evaluated on the task of HEp-2 classification. The proposed framework follows recent trends on feature encoding based on residuals.