Offline Bengali Writer Verification by PDF-CNN and Siamese Net
Authors:
- Chandranath AdakCAChandranath Adak
- Simone MarinaiSMSimone Marinai
- Bidyut B. ChaudhuriBCBidyut B. Chaudhuri
- Michael BlumensteinMBMichael Blumenstein
- 1 April 2018
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 381-386
Abstract
Automated handwriting analysis is a popular area of research owing to the variation of writing patterns. In this research area, writer verification is one of the most challenging branches, having direct impact on biometrics and forensics. In this paper, we deal with offline writer verification on complex handwriting patterns. Therefore, we choose a relatively complex script, i.e., Indic Abugida script Bengali (or, Bangla) containing more than 250 compound characters. From a handwritten sample, the probability distribution functions (PDFs) of some handcrafted features are obtained and input to a convolutional neural network (CNN). For such a CNN architecture, we coin the term "PDFCNN", where handcrafted feature PDFs are hybridized with auto-derived CNN features. Such hybrid features are then fed into a Siamese neural network for writer verification. The experiments are performed on a Bengali offline handwritten dataset of 100 writers. Our system achieves encouraging results, which sometimes exceed the results of state-of-the-art techniques on writer verification.Keywords
This publication has 26 references indexed in Scilit:
- Real-time single-pass connected components analysis algorithmEURASIP Journal on Image and Video Processing, 2013
- Writer verification using texture-based featuresInternational Journal on Document Analysis and Recognition (IJDAR), 2011
- Self-Organizing Maps for Clustering in Document Image Analysis2008
- BIOMETRIC WRITER IDENTIFICATION: FEATURE ANALYSIS AND CLASSIFICATIONInternational Journal of Pattern Recognition and Artificial Intelligence, 2006
- A writer identification and verification systemPattern Recognition Letters, 2005
- Touching numeral segmentation using water reservoir conceptPattern Recognition Letters, 2003
- A complete printed Bangla OCR systemPattern Recognition, 1998
- SIGNATURE VERIFICATION USING A “SIAMESE” TIME DELAY NEURAL NETWORKInternational Journal of Pattern Recognition and Artificial Intelligence, 1993
- Automatic signature verification and writer identification — the state of the artPattern Recognition, 1989
- IX. On the problem of the most efficient tests of statistical hypothesesPhilosophical Transactions of the Royal Society A, 1933