Colour laser printer identification using halftone texture fingerprint

Abstract

Numerous forgeries are made by precise and fast colour laser printers, and they have the ability to cause severe harm to society. To prevent such forgeries, printer identification can be employed as a countermeasure. A new method is presented to identify colour laser printers using halftone texture fingerprints. The method uses images photographed without an additional close-up lens as input images, and halftone texture fingerprints are extracted in the curvelet transform domain. The extracted halftone texture fingerprint is used in correlation-based detection, and the colour laser printer of the most similar known halftone texture fingerprint is determined as the source colour laser printer. Experiments are performed on five colour laser printers and the performance is compared with existing methods. Experimental results show that the method overcomes the limitations of existing methods.

Learning deep features for source color laser printer identification based on cascaded learning

“…While the printer identification methods mentioned above use scanned images, Kim and Lee [13] presented a method Kim's method can identify a source color laser printer with photographed images; however, it required an additional close-up lens to acquire useful input images. Therefore, they suggested a method using a halftone texture fingerprint extracted from photographed images [14]. The method proposed in [14] used photographed images that do not require an additional close-up lens as input images.…”

“…Source color laser printer identification techniques that use scanned images as input images could not identify the source printer of photographed input images, for either close-up photographed images or normally photographed images. To overcome this limitation of the color laser printer identification techniques, Kim and Lee [13] presented a method that uses halftone texture features of photographed images. They extracted texture features from close-up photographed halftone images and used the extracted features to train and test the SVM.…”

“…In [13], they used three sorts of halftone texture features: printing angle, printing resolution, and detail texture. However, it is difficult to extract the detail texture feature from normally photographed images.…”

“…However, it is difficult to extract the detail texture feature from normally photographed images. The method presented in [13] was not able to analyze normally photographed halftone images taken from mobile devices. Therefore, they extracted the halftone texture fingerprint from normally photographed images [14], and they used it in source color laser printer identification.…”

Application of Adobe® Photoshop® CC 2018 for identifying color laser printer source of Xerox® brand

“…Therefore, the determination of color laser printers source seized in forgery crimes became one of the most challenges in digital forensic labs today. In the last decade, digital forensic labs have introduced excess approaches to try catching a source of a defendant color laser printer such as, halftone analysis (Kim & Lee, 2014), statistical analysis of discrete wavelet transform (J. Choi et al, 2009) and noisy features analysis (J. H. Choi, Lee, & Lee, 2013), etc.…”