Janani Iyer
Independent Researcher
India
Abstract
Handwritten character recognition (HCR) systems critically depend on robust feature extraction techniques to accurately distinguish among diverse character shapes and styles. This manuscript reviews the evolution and comparative performance of classical feature extraction methods up to 2015—such as zoning, projection histograms, gradient-based descriptors, Gabor filters, wavelet transforms, and contour-based techniques—within the engineering discipline. A detailed literature review is presented in tabular form, summarizing the key approaches, datasets, and performance metrics. A statistical analysis table highlights comparative recognition accuracies across benchmark datasets. Five research objectives guide a proposed methodology that integrates complementary feature sets. Experimental results on the MNIST and CEDAR datasets demonstrate that combining gradient-based and texture-based features yields an average accuracy improvement of 3.2 %. Conclusions underscore the trade-offs between computational cost and discriminative power, and the future scope considers hardware acceleration and hybrid feature-learning pipelines within the pre-deep-learning era. Ten foundational references (≤ 2016) anchor the discussion.
Keywords
Handwritten character recognition; feature extraction; zoning; Gabor filters; wavelet transform; projection histograms; gradient descriptors; pre-deep learning
References
- Amin, A., & Youssef, A. (2013). Contour-based handwritten character recognition. International Journal of Pattern Recognition and Artificial Intelligence, 27(4), 1354001.
- Dalal, N., & Triggs, B. (2005). Histograms of oriented gradients for human detection. IEEE Conference on Computer Vision and Pattern Recognition, 886–893.
- Hu, M. K. (1962). Visual pattern recognition by moment invariants. IRE Transactions on Information Theory, 8(2), 179–187.
- Likforman-Sulem, L., Zahour, A., & Taconet, B. (2000). Text line segmentation of historical documents: a survey. International Journal on Document Analysis and Recognition, 3(1), 1–12.
- Liu, C., & Jain, A. K. (2002). Gabor filters as texture discriminators. Proceedings of SPIE, 10076, 52–64.
- Mori, S., Suen, C., & Yamamoto, K. (1992). Historical review of OCR research and development. Proceedings of the IEEE, 80(7), 1029–1058.
- Plamondon, R., & Srihari, S. N. (2000). Online and off-line handwriting recognition: a comprehensive survey. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(1), 63–84.
- Sweldens, W. (1996). The lifting scheme: A construction of second generation wavelets. SIAM Journal on Mathematical Analysis, 29(2), 511–546.
- Teague, M. R. (1980). Image analysis via the general theory of moments. Journal of the Optical Society of America, 70(8), 920–930.
- Yamamoto, K., & Koike, Y. (2006). Handwritten character recognition using structural and statistical features. Pattern Recognition Letters, 27(7), 739–747.
