Abstract
There are many scenarios where we wish to imitate a specific author’s pen-on-paper handwriting style. Rendering new text in someone’s handwriting is difficult because natural handwriting is highly variable, yet follows both intentional and involuntary structure that makes a person’s style self-consistent. The variability means that naive example-based texture synthesis can be conspicuously repetitive.
We propose an algorithm that renders a desired input string in an author’s handwriting. An annotated sample of the author’s handwriting is required; the system is flexible enough that historical documents can usually be used with only a little extra effort. Experiments show that our glyph-centric approach, with learned parameters for spacing, line thickness, and pressure, produces novel images of handwriting that look hand-made to casual observers, even when printed on paper.
Supplemental Material
Available for Download
Supplemental movie, appendix, image and software files for, My Text in Your Handwriting
- R. M. Alexander. 1984. The gaits of bipedal and quadrupedal animals. Robotics Research 3, 2, 49--59.Google ScholarCross Ref
- J. Andre and B. Borghi. 1990. Dynamic fonts. Raster Imaging and Digital Typography 198--204.Google Scholar
- A. Antonacopoulos, B. Gatos, and D. Bridson. 2007. Page segmentation competition. Document Analysis and Recognition 2, 1279--1283. Google ScholarDigital Library
- R. Bellman. 1952. On the theory of dynamic programming. PNAS 38, 8, 716--719.Google ScholarCross Ref
- S. Belongie and J. Malik. 2000. Matching with shape contexts. Content-Based Access of Image and Video Libraries 81--105.Google Scholar
- H. Blum. 1967. A transformation for extracting new descriptors of shape. Models for the Perception of Speech and Visual Form.Google Scholar
- N. Bonneel, M. van de Panne, S. Lefebvre, and G. Drettakis. 2010. Proxy-guided texture synthesis for rendering natural scenes. Proc. Vision Modeling and Visualization.Google Scholar
- F. L. Bookstein. 1989. Principal warps: Thin-plate splines and the decomposition of deformations. PAMI 11, 6, 567--585. Google ScholarDigital Library
- Y. Boykov and V. Kolmogorov. 2004. An experimental comparison of min-cut/max-flow algorithms for energy minimization in vision. PAMI 26, 9, 1124--1137. Google ScholarDigital Library
- Y. Y. Boykov and M.-P. Jolly. 2001. Interactive graph cuts for optimal boundary & region segmentation of objects in N-D images. ICCV 105--112.Google Scholar
- M. Brand and A. Hertzmann. 2000. Style machines. SIGGRAPH 183--192. Google ScholarDigital Library
- L. Breiman. 2001. Random forests. Machine Learning 45, 1, 5--32. Google ScholarDigital Library
- N. D. F. Campbell and J. Kautz. 2014. Learning a manifold of fonts. TOG 33, 4. Google ScholarDigital Library
- W.-D. Chang and J. Shin. 2012. A statistical handwriting model for style-preserving and variable character synthesis. Document Analysis and Recognition 15, 1, 1--19. Google ScholarDigital Library
- H. Choi, S. J. Cho, and J. H. Kim. 2004. Writer dependent online handwriting generation with bayesian networks. Frontiers in Handwriting Recognition 130--135. Google ScholarDigital Library
- S. Chowdhury, S. Das, D. Roy, U. Sarkar, and B. B. Chaudhuri. 2009. A complete method of personal handwriting synthesis. Advances in Graphonomics 250--253.Google Scholar
- D. Comaniciu and P. Meer. 2002. Mean shift: A robust approach toward feature space analysis. PAMI 24, 5, 603--619. Google ScholarDigital Library
- S. D. Connell and A. K. Jain. 2002. Writer adaptation for online handwriting recognition. PAMI 24, 3, 329--346. Google ScholarDigital Library
- S. Cooper, A. Hertzmann, and Z. Popovic. 2007. Active learning for real-time motion controllers. TOG 26, 3. Google ScholarDigital Library
- S. Correll. 2000. Graphite: An extensible rendering engine for complex writing systems. Unicode Conference 17.Google Scholar
- A. Criminisi and J. Shotton. 2013. Decision Forests for Computer Vision and Medical Image Analysis. Springer. Google ScholarDigital Library
- N. Dalal and B. Triggs. 2005. Histograms of orientated gradients for human detection. CVPR 886--893. Google ScholarDigital Library
- L. Devroye and M. McDougall. 1995. Random fonts for the simulation of handwriting. Electronic Publishing 8, 4, 281--294.Google Scholar
- L. Dinges, A. Al-Hamadi, and M. Elzobi. 2013. An approach for arabic handwriting synthesis based on active shape models. Document Analysis and Recognition 1260--1264. Google ScholarDigital Library
- A. C. Doyle. 1901. Letter from Arthur Conan Doyle to Herbert Greenhough Smith. Retrieved from https://commons.wikimedia.org/wiki/File:Letter_from_Arthur_Conan_Doyle_to_Herbert_Greenhough_Smith.jpg, Creative Commons Attribution-Share Alike 2.0, Toronto Public Library. Accessed: 2016-02-09.Google Scholar
- D. S. Ebert, F. K. Musgrave, D. Peachey, K. Perlin, and S. Worley. 2002. Texturing and Modeling: A Procedural Approach. Morgan Kaufmann. Google ScholarDigital Library
- A. A. Efros and W. T. Freeman. 2001. Image quilting for texture synthesis and transfer. SIGGRAPH 28, 341--346. Google ScholarDigital Library
- A. A. Efros and T. K. Leung. 1999. Texture synthesis by non-parametric sampling. ICCV 2, 1033--1038. Google ScholarDigital Library
- A. Einstein. 1956. Investigations on the theory of the Brownian movement. Courier Dover Publications.Google Scholar
- Y. Elarian, R. Abdel-Aal, I. Ahmad, M. T. Parvez, and A. Zidouri. 2014. Handwriting synthesis: Classifications and techniques. IJDAR 17, 455--469. Google ScholarDigital Library
- Y. Elarian, I. Ahmad, S. Awaida, W. G. Al-Khatib, and A. Zidouri. 2015. An arabic handwriting synthesis system. Pattern Recognition 48, 3, 849--861. Google ScholarDigital Library
- Y. S. Elarian, H. A. Al-Muhtaseb, and L. M. Ghouti. 2011. Arabic handwriting synthesis. Frontiers in Arabic Handwriting Recognition.Google Scholar
- L. R. Ford and D. R. Fulkerson. 1956. Maximal flow through a network. Canadian Journal of Mathematics 8, 399--404.Google ScholarCross Ref
- K. Fukunaga and L. D. Hostetler. 1975. The estimation of the gradient of a density function, with applications in pattern recognition. Trans. Information Theory 21, 32--40. Google ScholarDigital Library
- G. Gangadhar, D. Joseph, and V. S. Chakravarthy. 2007. An oscillatory neuromotor model of handwriting generation. Document Analysis and Recognition 10, 2, 69--84. Google ScholarDigital Library
- R. Garner. 2005. Post-it note persuasion: A sticky influence. J. Consumer Psychology 15, 230--237.Google ScholarCross Ref
- A. Graves. 2013. Generating sequences with recurrent neural networks. arXiv Preprint.Google Scholar
- W. Guerfali and R. Plamondon. 1995. The delta lognormal theory for the generation and modeling of cursive characters. Document Analysis and Recognition 3, 1. Google ScholarDigital Library
- G. Gupta and A. Mccabe. 1997. A Review of Dynamic Handwritten Signature Verification. Technical Report. James Cook University.Google Scholar
- I. Guyon. 1996. Handwriting synthesis from handwritten glyphs. Frontiers of Handwriting Recognition 309--312.Google Scholar
- T. S. F. Haines and T. Xiang. 2014. Active rare class discovery and classification using Dirichlet processes. IJCV 106, 3, 315--331. Google ScholarDigital Library
- G. Hinton and V. Nair. 2005. Inferring motor programs from images of handwritten digits. NIPS 515--522.Google Scholar
- T. K. Ho. 1995. Random decision forests. Proc. Document Analysis and Recognition 1, 278--282. Google ScholarDigital Library
- W. Huang, Z. Lin, J. Yang, and J. Wang. 2013. Text localization in natural images using stroke feature transform and text covariance descriptors. ICCV. Google ScholarDigital Library
- Charles Jacobs, Wil Li, Evan Schrier, David Bargeron, and David Salesin. 2004. Adaptive document layout. Commun. ACM 47, 8, 60--66. Google ScholarDigital Library
- R. E. Kalman. 1960. A new approach to linear filtering and prediction problems. ASME--Basic Engineering 82, 35--45.Google ScholarCross Ref
- D. E. Knuth. 1979. METAFONT: A System for Alphabet Design. Technical Report. Stanford University. Google ScholarDigital Library
- V. Kwatra, A. Schödl, I. Essa, G. Turk, and A. Bobick. 2003. Graphcut textures: Image and video synthesis using graph cuts. TOG 22, 3, 277--286. Google ScholarDigital Library
- Hyunjun Lee, Sungtae Kwon, and Seungyong Lee. 2006. Real-time pencil rendering. In NPAR. 37--45. Google ScholarDigital Library
- J. P. Lewis. 1984. Texture synthesis for digital painting. Computer Graphics 18, 3. Google ScholarDigital Library
- Z. Lin and L. Wan. 2007. Style-preserving english handwriting synthesis. Pattern Recognition 40, 2097--2109. Google ScholarDigital Library
- D. G. Lowe. 2004. Distinctive image features from scale-invariant keypoints. IJCV. Google ScholarDigital Library
- J. Lu, F. Yu, A. Finkelstein, and S. DiVerdi. 2012. HelpingHand: Example-based stroke stylization. SIGGRAPH 31, 4. Google ScholarDigital Library
- J. Mao and K. M. Mohiuddin. 1997. Improving OCR performance using character degradation models and boosting algorithm. Pattern Recognition Letters 18, 1415--1419. Google ScholarDigital Library
- B. Marcelo, G. Sapiro, V. Caselles, and C. Ballester. 2000. Image inpainting. SIGGRAPH 27, 417--424. Google ScholarDigital Library
- U.-V. Marti and H. Bunke. 2002. Using a statistical language model to improve the performance of an HMM-based cursive handwriting recognition systems. PRAI 15.Google Scholar
- U. Mohammed, S. J. D. Prince, and J. Kautz. 2009. Visio-lization: Generating novel facial images. TOG 28, 3. Google ScholarDigital Library
- M. Mori, A. Suzuki, A. Shio, and S. Ohtsuka. 2000. Generating new samples from handwritten numerals based on point correspondence. IWFHR 7, 281--290.Google Scholar
- S. Mueller, N. Huebel, M. Waibel, and R. D. Andrea. 2013. Robotic calligraphy - learning how to write single strokes of Chinese and Japanese characters. IROS 1734--1739.Google Scholar
- P. Norvig. 2014. Letter Frequencies for Scrabble. http://norvig.com/scrabble-letter-scores.html. Accessed: 2014-05-03.Google Scholar
- S. J. Pan and Q. Yang. 2010. A survey on transfer learning. Knowledge and Data Engineering 22, 10, 1345--1359. Google ScholarDigital Library
- G. Papandreou and A. L. Yuille. 2011. Perturb-and-MAP random fields: Using discrete optimization to learn and sample from energy models. ICCV 193--200. Google ScholarDigital Library
- S. N. Parizi, A. Vedaldi, A. Zisserman, and P. Felzenszwalb. 2014. Automatic discovery and optimization of parts for image classification. arXiv Preprint, arXiv:1412.6598.Google Scholar
- V. Pervouchine and G. Leedham. 2007. Extraction and analysis of forensic document examiner features used for writer identification. Pattern Recognition 40, 3, 1004--1013. Google ScholarDigital Library
- R. Plamondon and S. N. Srihari. 2000. On-line and off-line handwriting recognition: A comprehensive survey. PAMI 22, 1, 63--84. Google ScholarDigital Library
- T. Porter and T. Duff. 1984. Compositing digital images. Computer Graphics 18, 253--259. Google ScholarDigital Library
- Javier Portilla and Eero P. Simoncelli. 2000. A parametric texture model based on joint statistics of complex wavelet coefficients. IJCV 40, 1, 49--70. Google ScholarDigital Library
- Project Gutenberg. 2014. Homepage. Retrieved from http://www.gutenberg.org/. Accessed: 2014-05-03.Google Scholar
- Tobias Reinhardt. 2014. My Script Font. Retrieved from http://myscriptfont.com. Accessed: 2014-05-03.Google Scholar
- C. Rother, V. Kolmogorov, and A. Blake. 2004. “GrabCut” interactive foreground extraction using iterated graph cuts. TOG 23, 3, 309--314. Google ScholarDigital Library
- Scrabble. 2014. History of Scrabble. Retrieved from http://www.scrabble-assoc.com/info/history.html. Accessed: 2014-05-03.Google Scholar
- A. Schödl, R. Szeliski, D. H. Salesin, and I. Essa. 2000. Video textures. SIGGRAPH 27, 489--498. Google ScholarDigital Library
- M. Schroder. 2001. Emotional speech synthesis: A review. Interspeech 561--564.Google Scholar
- M. Shilman, D. S. Tan, and P. Simard. 2006. CueTIP: A mixed-initiative interface for correcting handwriting errors. User Interface Software and Technology 323--332. Google ScholarDigital Library
- S. N. Srihari, S.-H. Cha, H. Arora, and S. Lee. 2002. Individuality of handwriting. Journal of Forensic Science 47, 4.Google ScholarCross Ref
- J. Sun, L. Yuan, J. Jia, and H.-Y. Shum. 2005. Image completion with structure propagation. TOG 24, 3, 861--868. Google ScholarDigital Library
- A. O. Thomas, A. Rusu, and V. Govindaraju. 2009. Synthetic handwritten CAPTCHAs. Pattern Recognition 42, 12, 3365--3373. Google ScholarDigital Library
- G. P. van Galen. 1991. Handwriting: Issues for a psychomotor theory. Human Movement Science 10, 165--191.Google ScholarCross Ref
- T. Varga and H. Bunke. 2003. Generation of synthetic training data for an HMM-based handwriting recognition system. Document Analysis and Recognition 618--622. Google ScholarDigital Library
- T. Varga and H. Bunke. 2004. Comparing natural and synthetic training data for off-line cursive handwriting recognition. Frontiers in Handwriting Recognition 9, 221--225. Google ScholarDigital Library
- L. Wan and Z. Lin. 2009. Signature sample synthesis. Encyclopedia of Biometrics 1205--1210.Google Scholar
- J. Wang, C. Wu, Y.-Q. Xu, and H.-Y. Shum. 2004. Combining shape and physical models for online cursive handwriting synthesis. Document Analysis and Recognition 7, 4, 219--227. Google ScholarDigital Library
- J. Wang, C. Wu, Y.-Q. Xu, H.-Y. Shum, and L. Ji. 2002. Learning-based cursive handwriting synthesis. Frontiers in Handwriting Recognition 8, 157--162. Google ScholarDigital Library
- Y. Wang, H. Wang, C. Pan, and L. Fang. 2008. Style preserving Chinese character synthesis based on hierarchical representation of character. Acoustics, Speech and Signal Processing 1097--1100.Google Scholar
- J. Warnock, C. Geschke, D. Brotz, E. Taft, and B. Paxton. 1984. PostScript. Retrieved from https://www.adobe.com/products/postscript/pdfs/postscript_is_20.pdf. Accessed: 2014-11-29.Google Scholar
- L.-Y. Wei, S. Lefebvre, V. Kwatra, and G. Turk. 2009. State of the art in example-based texture synthesis. Eurographics 93--117.Google Scholar
- L. Werner. 1999. Getting Java ready for the world: A brief history of IBM and Sun’s internationalization efforts. Retrieved from http://www.icu-project.org/docs/papers/history_of_java_internationalization.html. Accessed: 2014-12-6.Google Scholar
- B. H. Williams, M. Toussaint, and A. J. Storkey. 2007. Modelling motion primitive and their timing in biologically executed movements. NIPS 1609--1616.Google Scholar
- George Williams. 2014. Font Forge. Retrieved from http://fontforge.org/.Google Scholar
- C. Xiong, D. Johnson, R. Xu, and J. J. Corso. 2012. Random forests for metric learning with implicit pairwise position dependence. Knowledge Discovery and Data Mining 18, 958--966. Google ScholarDigital Library
- T. Y. Zhang and C. Y. Suen. 1984. A fast parallel algorithm for thinning digital patterns. Commun. ACM 27, 3, 236--239. Google ScholarDigital Library
- Y. Zheng and D. Doermann. 2005. Handwriting matching and its application to handwriting synthesis. Document Analysis and Recognition 8, 861--865. Google ScholarDigital Library
- C. L. Zitnick. 2013. Handwriting beautification using token means. TOG 32, 4. Google ScholarDigital Library
Recommendations
Automatic generation of large-scale handwriting fonts via style learning
SA '16: SIGGRAPH ASIA 2016 Technical BriefsGenerating personal handwriting fonts with large amounts of characters is a boring and time-consuming task. Take Chinese fonts as an example, the official standard GB18030-2000 for commercial font products contains 27533 simplified Chinese characters. ...
Style-preserving English handwriting synthesis
This paper presents a novel and effective approach to synthesize English handwriting in the user's writing style. We select the most important features that depict the handwriting style, including character glyph, size, slant, and pressure, special ...
Text entry performance of state of the art unconstrained handwriting recognition: a longitudinal user study
CHI '09: Proceedings of the SIGCHI Conference on Human Factors in Computing SystemsWe report on a longitudinal study of unconstrained handwriting recognition performance. After 250 minutes of practice, participants had a mean text entry rate of 24.1 wpm. For the first four hours of usage, entry and error rates of handwriting ...
Comments