Patch Autocorrelation Features for Optical Character Recognition

Radu Tudor Ionescu, Andreea-Lavinia Popescu, Dan Popescu


The autocorrelation is often used in signal processing as a tool for finding repeating patterns in a signal. In image processing, there are various image analysis techniques that use the autocorrelation of an image for a broad range of applications from texture analysis to grain density estimation. In this paper, a novel approach of capturing the autocorrelation of an image is proposed. More precisely, the autocorrelation is recorded in a set of features obtained by comparing pairs of patches from an image. Each feature stores the euclidean distance between a particular pair of patches. Although patches contain contextual information and have advantages in terms of generalization, most of the patch-based techinques used in image processing are heavy to compute with current machines. Therefore, patches are selected using a dense grid over the image to reduce the number of features. This approach is termed Patch Autocorrelation Features (PAF). The proposed approach is evaluated in a series of handwritten digit recognition experiments using the popular MNIST data set. The Patch Autocorrelation Features are compared with the euclidean distance using two classification systems, namely the k-Nearest Neighbors and Support Vector Machines. The empirical results show that the feature map proposed in this work is always better than a feature representation based on raw pixel values, in terms of accuracy. Furthermore, the results obtained with PAF are comparable to other state of the art methods.


  1. Agarwal, S. and Roth, D. (2002). Learning a Sparse Representation for Object Detection. Proceedings of ECCV, pages 113-127.
  2. Barnes, C., Goldman, D. B., Shechtman, E., and Finkelstein, A. (2011). The PatchMatch Randomized Matching Algorithm for Image Manipulation. Communications of the ACM, 54(11):103-110.
  3. Belongie, S., Malik, J., and Puzicha, J. (2002). Shape matching and object recognition using shape contexts. IEEE Transactions on Pattern Analysis and Machine Intelligence, 24(4):509-522.
  4. Brochard, J., Khoudeir, M., and Augereau, B. (2001). Invariant feature extraction for 3D texture analysis using the autocorrelation function. Pattern Recognition Letters, 22(6-7):759-768.
  5. Cho, T. S., Avidan, S., and Freeman, W. T. (2010). The patch transform. PAMI, 32(8):1489-1501.
  6. Ciresan, D. C., Meier, U., and Schmidhuber, J. (2012). Multi-column Deep Neural Networks for Image Classification. Proceedings of CVPR, pages 3642-3649.
  7. Cortes, C. and Vapnik, V. (1995). Support-vector networks. Machine Learning, 20(3):273-297.
  8. DeCoste, D. and Schölkopf, B. (2002). Training Invariant Support Vector Machines. Machine Learning, 46(1- 3):161-190.
  9. Deselaers, T., Keyser, D., and Ney, H. (2005). Discriminative Training for Object Recognition using Image Patches. Proceedings of CVPR, pages 157-162.
  10. Dinu, L. P., Ionescu, R., and Popescu, M. (2012). Local Patch Dissimilarity for Images. Proceedings of ICONIP, 7663:117-126.
  11. Efros, A. A. and Freeman, W. T. (2001). Image quilting for texture synthesis and transfer. Proceedings of SIGGRAPH 7801, pages 341-346.
  12. Guo, G. and Dyer, C. R. (2007). Patch-based Image Correlation with Rapid Filtering. Proceedings of CVPR.
  13. Horikawa, Y. (2004). Use of Autocorrelation Kernels in Kernel Canonical Correlation Analysis for Texture Classification. Proceedings of ICONIP, 3316:1235- 1240.
  14. Horikawa-2004 (2004). Comparison of support vector machines with autocorrelation kernels for invariant texture classification. Proceedings of ICPR, 1:660-663.
  15. Ionescu, R. T. and Popescu, M. (2014). PQ kernel: a rank correlation kernel for visual word histograms. Pattern Recognition Letters.
  16. Kégl, B. and Busa-Fekete, R. (2009). Boosting Products of Base Classifiers. Proceedings of ICML, pages 497- 504.
  17. Keysers, D., Deselaers, T., Gollan, C., and Ney, H. (2007). Deformation Models for Image Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29(8):1422-1435.
  18. LeCun, Y., Bottou, L., Bengio, Y., and Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278-2324.
  19. LeCun, Y., Jackel, L. D., Boser, B., Denker, J. S., Graf, H. P., Guyon, I., Henderson, D., Howard, R. E., and Hubbard, W. (1989). Handwritten digit recognition: Applications of neural net chips and automatic learning. IEEE Communications, pages 41-46.
  20. Paredes, R., Prez-Cortes, J., Juan, A., and Vidal, E. (2001). Local Representations and a Direct Voting Scheme for Face Recognition. Proceedings of Workshop on Pattern Recognition in Information Systems, pages 71- 79.
  21. Popovici, V. and Thiran, J. (2001). Higher order autocorrelations for pattern classification. Proceedings of ICIP, 3:724-727.
  22. Simard, P., LeCun, Y., Denker, J. S., and Victorri, B. (1996). Transformation Invariance in Pattern Recognition, Tangent Distance and Tangent Propagation. Neural Networks: Tricks of the Trade.
  23. Srihari, S. N. (1992). High-performance reading machines. Proceedings of the IEEE (Special issue on Optical Character Recognition), 80(7):1120-1132.
  24. Suen, C. Y., Nadal, C., Legault, R., Mai, T. A., and Lam, L. (1992). Computer recognition of unconstrained handwritten numerals. Proceedings of the IEEE (Special issue on Optical Character Recognition), 80(7):1162- 1180.
  25. Teow, L.-N. and Loe, K.-F. (2002). Robust visionbased features and classification schemes for off-line handwritten digit recognition. Pattern Recognition, 35(11):2355-2364.
  26. Toyoda, T. and Hasegawa, O. (2007). Extension of higher order local autocorrelation features. Pattern Recognition, 40(5):1466-1473.
  27. Wilder, K. J. (1998). Decision tree algorithms for handwritten digit recognition. Electronic Doctoral Dissertations for UMass Amherst.

Paper Citation

in Harvard Style

Ionescu R., Popescu A. and Popescu D. (2015). Patch Autocorrelation Features for Optical Character Recognition . In Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2015) ISBN 978-989-758-089-5, pages 419-426. DOI: 10.5220/0005289804190426

in Bibtex Style

author={Radu Tudor Ionescu and Andreea-Lavinia Popescu and Dan Popescu},
title={Patch Autocorrelation Features for Optical Character Recognition},
booktitle={Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2015)},

in EndNote Style

JO - Proceedings of the 10th International Conference on Computer Vision Theory and Applications - Volume 1: VISAPP, (VISIGRAPP 2015)
TI - Patch Autocorrelation Features for Optical Character Recognition
SN - 978-989-758-089-5
AU - Ionescu R.
AU - Popescu A.
AU - Popescu D.
PY - 2015
SP - 419
EP - 426
DO - 10.5220/0005289804190426