A New Method for Detecting Altered Text in Document Images

Nandanwar, Lokesh; Shivakumara, Palaiahnakote; Pal, Umapada; Lü, Tong; Lopresti, Daniel; Seraogi, Bhagesh; Chaudhuri, B. B.

doi:10.1007/978-3-030-59830-3_8

Cited by 6 publications

(9 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For comparison purposes, we tested a number of other methods for related problems. These include Shivakumara et al [24] which was developed for forged IMEI number detection; Wang et al [23], which was developed for printer identification; Nandanwar et al [30], which was developed for altered text detection in document images. Since the objective of these three methods is the same as our proposed method, they make good choices for a comparative study.…”

Section: Resultsmentioning

confidence: 99%

“…This is because the method is developed for studying compression errors caused by blur and distortion. However, it is worst for the insertion data compared to [23, 25, 30, 33]. This is because insertion operations do not introduce much blur, rather they change character shapes.…”

Section: Resultsmentioning

confidence: 99%

“…This is because insertion operations do not introduce much blur, rather they change character shapes. The method in [30] achieves the best F ‐score compared to the proposed method and other existing methods for the insertion data because the features proposed in [30] are robust, which work based on the quality of the images but not extracting specific features. The method [25] has the best average classification rate for the insertion data.…”

Section: Resultsmentioning

confidence: 99%

“…The results of the proposed and existing methods are reported in Table 4, where it can be seen that the proposed method is slightly better than the existing methods in terms of average classification rate and F ‐score. When we look at the results of the existing methods, the method [30] scores the highest results compared to the other existing methods. This is because the features extracted are based on the quality of the images, while the other existing methods use specific features based on text information.…”

Section: Resultsmentioning

confidence: 99%

“…Then review of existing work shows that the methods focus on one type of image, whereas the proposed work considers different types of images in this paper. Nandanwar et al [30] proposed method for detecting altered text in document images. Their approach splits the Discrete Cosine Transform (DCT) into positive and negative coefficients and reconstructs images for the respective positive and negative coefficients.…”

Section: Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Forged text detection in video, scene, and document images

et al. 2020

Self Cite

View full text Add to dashboard Cite

Rapid advances in artificial intelligence have made it possible to produce forgeries good enough to fool an average user. As a result, there is growing interest in developing robust methods to counter such forgeries. This study presents a new Fourier spectrum‐based method for detecting forged text in video images. The authors' premise is that brightness distribution and the spectrum shape exhibit irregular patterns (inconsistencies) for forged text, while appearing more regular for original text. The method divides the spectrum of an input image into sectors and tracks to highlight these effects. Specifically, positive and negative coefficients for sectors and tracks are extracted to quantify the brightness distribution. Variations in the shape of the spectrum are analysed by determining the angular relationship between the principal axes and the sectors/tracks of the spectrum. Next, it combines these two features to detect forged text in the images of IMEI (International Mobile Equipment Identity) numbers and document. For evaluation, the following datasets are used: own video dataset and standard datasets, namely, IMEI number, ICPR 2018 Fraud Document Contest, and a natural scene text dataset. Experimental results show that the proposed method outperforms existing methods in terms of average classification rate and F‐score.

show abstract

Section: Resultsmentioning

confidence: 99%