Data-Driven Feature Characterization Techniques for Laser Printer Attribution

Ferreira, Anselmo; Bondi, Luca; Baroffio, Luca; Bestagini, Paolo; Huang, Jiwu; Santos, Jefersson A. dos; Tubaro, Stefano; Rocha, Anderson

doi:10.1109/tifs.2017.2692722

Cited by 54 publications

(100 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• The data-driven method CNN-{S raw , S med , S avg } a,e also performs well when it is trained using Cambria font and tested on the same font (100% accuracy, 3 rd row -2 nd column of Thus, the same font results in Table IV concur with published literature [2], [10] and clearly demonstrate that the existing methods do very well when the font of letters present in testing data is also present in training data. However, these methods are not suitable for the scenarios where font of letters in testing data is different from that in training data.…”

Section: ) Same Font Experimentssupporting

confidence: 79%

“…First, we evaluate the performance of our proposed PSLTD based method on DB1 [7] which is publicly available. This dataset contains characters of mixed font types on the same page so we perform (i) letter specific (using all occurrences of letter 'e') evaluation similar to [2] and [7] and (ii) universal letter evaluation (using all connected components as suggested in [10]).…”

Section: B Experiments On Db1mentioning

confidence: 99%

“…Under these settings, the proposed method (denoted as PSLTD e ) achieves an average page-level classification accuracy of 98.92% with a standard deviation of ±0.43 which is a gain of 1.8% over the best reported accuracy only on e's, using CC-RS-LTrP-PoEP e [10] (Table I). In addition, the state-of-the-art CNN based method (CNN-{S raw , S med , S avg } a,e [2]) gives an average page-level classification accuracy of 97.33% using all occurrences of 'e' and 'a' images along with their median and average residuals. Thus, the proposed PSLTD e outperforms existing methods by correctly classifying, on an average, about 9 more pages out of 592 test pages.…”

Section: ) Letter Specific Evaluationmentioning

confidence: 99%

“…• Evaluation of the proposed system comprehensively on the publicly available dataset [2], [7], showing that, given sufficient amount of train data, it outperforms existing methods.…”

Section: Introductionmentioning

confidence: 99%

“…{S raw , S med , S avg } a,e[2], CTGF-GLCM-MD-MS e[2] and the proposed PSLTD on the existing dataset[7] corresponding to 5 × 2 cross-validation folds of[2].…”

mentioning

confidence: 99%

See 4 more Smart Citations

Source Printer Classification Using Printer Specific Local Texture Descriptor

Joshi

Khanna

2020

IEEE Trans.Inform.Forensic Secur.

View full text Add to dashboard Cite

The knowledge of source printer can help in printed text document authentication, copyright ownership, and provide important clues about the author of a fraudulent document along with his/her potential means and motives. Development of automated systems for classifying printed documents based on their source printer, using image processing techniques, is gaining lot of attention in multimedia forensics. Currently, state-of-the-art systems require that the font of letters present in test documents of unknown origin must be available in those used for training the classifier. In this work, we attempt to take the first step towards overcoming this limitation. Specifically, we introduce a novel printer specific local texture descriptor. The highlight of our technique is the use of encoding and regrouping strategy based on small linear-shaped structures composed of pixels having similar intensity and gradient. The results of experiments performed on two separate datasets show that: 1) on a publicly available dataset, the proposed method outperforms state-of-the-art algorithms for characters printed in the same font, and 2) on another dataset 1 having documents printed in four different fonts, the proposed method correctly classifies all test samples when sufficient training data is available in same font setup. In addition, it outperforms state-of-the-art methods for cross font experiments. Moreover, it reduces the confusion between the printers of same brand and model.

show abstract

Section: ) Same Font Experimentssupporting

confidence: 79%

Section: B Experiments On Db1mentioning

confidence: 99%

Section: ) Letter Specific Evaluationmentioning

confidence: 99%

“…• Evaluation of the proposed system comprehensively on the publicly available dataset [2], [7], showing that, given sufficient amount of train data, it outperforms existing methods.…”

Section: Introductionmentioning

confidence: 99%

“…{S raw , S med , S avg } a,e[2], CTGF-GLCM-MD-MS e[2] and the proposed PSLTD on the existing dataset[7] corresponding to 5 × 2 cross-validation folds of[2].…”

mentioning

confidence: 99%

See 3 more Smart Citations