Projection pursuit and PCA associated with near and middle infrared hyperspectral images to investigate forensic cases of fraudulent documents

详细信息    查看全文

• 作者：José ; Francielson Queiroz Pereira^a ; Carolina S. Silva^a ; André ; Braz^b ; Maria Fernanda Pimentel^b ; ^mfp@ufpe.br ; Ricardo Saldanha Honorato^c ; Celio Pasquini^d ; Peter D. Wentzell^e
• 关键词：Hyperspectral image ; Projection pursuit ; Documents ; Forensic ; Principal component analysis
• 刊名：Microchemical Journal
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：130
• 期：Complete
• 页码：412-419
• 全文大小：1342 K
• 卷排序：130

文摘

Hyperspectral infrared images and PCA-PP techniques for fraud detection. ROI selection and PCA histograms improved spectral quality and discrimination. PP showed better discrimination results of black pen inks than PCA. PCA and PP combined with HIS-NIR and HIS-MIR solved 90% blind samples. Methodology developed was effective for detection of fraudulent documents.