Abstract
New Tai Lue is widely used in Southwest China and Southeast Asia. Hence, it is important to study related handwritten character recognition. Considering the many similar characters in handwritten New Tai Lue, this paper proposes an offline handwritten New Tai Lue character recognition method based on convolutional prior features and deep variationally sparse Gaussian process (DVSGP) modeling. An offline handwritten database is constructed, a convolutional neural network is trained to extract the convolutional features of New Tai Lue character images as prior features, and a DVSGP model is built. The extracted features are input into the DVSGP model to construct a recognition model. The experimental results show that the accuracy of the model is 97.67% and that the precision, recall, and F1-score are 0.9769, 0.9767, and 0.9767, respectively, which are better than those of other methods. The proposed method also achieves high accuracy on the MNIST recognition task, verifying its universal applicability.
- [1] . 2020. Classification of ancient handwritten Tamil characters on palm leaf inscription using modified adaptive backpropagation neural network with GLCM features. ACM Trans. Asian Low-Resour. Lang. Inf. Process. 19, 6 (November 2020), 1–24. https://doi.org/10.1145/3406209 Google ScholarDigital Library
- [2] . 2018. Drawing and recognizing Chinese characters with recurrent neural network. IEEE Trans. Pattern Anal. Mach. Intell. 40, 4 (April 2018), 849–862. 10.1109/TPAMI.2017.2695539Google ScholarCross Ref
- [3] . 2018. Online handwritten character recognition of New Tai Lue based on online random forests Adv. Intell. Sys. Comput. 686, (August 2018), 426–434. https://doi.org/10.1007/978-3-319-69096-4_59Google Scholar
- [4] . 2019. Handwritten Manipuri Meetei-Mayek classification using convolutional neural network. ACM Trans. Asian Low-Reso. 18, 4,
Article 35 (August 2019), 1–23. https://doi.org/10.1145/33094977 Google ScholarDigital Library - [5] . 2018. A deep convolutional generative adversarial networks (DCGANs)-based semi-supervised method for object recognition in synthetic aperture radar (SAR) images. Remote Sens-Basel. 10, 6 (2018), 846. https://doi.org/10.3390/rs10060846Google ScholarCross Ref
- [6] . 2013. Deep Gaussian processes, arXiv:1211.0358 12, (March 2013), 207–215.Google Scholar
- [7] . 2018. Efficient global optimization using deep Gaussian processes. IEEE Congr. Evol. Comput., CEC - Conf. Proc. (July 2018), 1–8. https://doi.org/10.1109/CEC.2018.8477946Google Scholar
- [8] . 2016. Mongolian handwriting character recognition based on convolutional neural network (CNN). Sixth Appl. Mech. Mater. 2016 (October 2016), 580–584. https://doi.org/10.2991/icimm-16.2016.105Google Scholar
- [9] . 2015. A dictionary learning and KPCA-based feature extraction method for off-line handwritten Tibetan character recognition. Optik 126, 23 (December 2015), 3795–3800. https://doi.org/10.1016/j.ijleo.2015.07.144Google ScholarCross Ref
- [10] . 2016. Printed New Tai Lue character recognition based on BP neural network. IEEE Int. Conf. Signal Image Process. ICSIP (August 2016), 339–342. https://doi.org/10.1109/SIPROCESS.2016.7888280Google Scholar
- [11] . 2017. Online handwritten character recognition of New Tai Lue based on online random forests. Adv. Intell. Sys. Comput. (November 2017), 70–74. https://doi.org/10.1007/978-3-319-69096-4_59Google Scholar
- [12] . 2013. Evaluation of weighted Fisher criteria for large category dimensionality reduction in application to Chinese handwriting recognition. Pattern Recognt. 46, 9 (September 2013), 2599–2611. https://doi.org/10.1016/j.patcog.2013.01.036 Google ScholarDigital Library
- [13] , 2016. Improving handwritten Chinese text recognition using neural network language models and convolution neural network shape models. Pattern Recognt. 65, (May 2017), 251–264. https://doi.org/10.1016/j.patcog.2016.12.026 Google ScholarDigital Library
- [14] . 2019. Recognizing online handwritten Chinese characters using RNNs with new computing architectures. Pattern Recognt. 93, (September 2019), 179–192. https://doi.org/10.1016/j.patcog.2019.04.015Google ScholarCross Ref
- [15] . 2017. Online and offline handwritten Chinese character recognition: A comprehensive study and new benchmark. arXiv:1606.05763. 61 (June 2017), 348–360. https://doi.org/10.1016/j.patcog.2016.08.005Google Scholar
- [16] . 2017. Printed Thai character segmentation and recognition. IEEE Int. Conf. Soft Comput. Mach. Intell., ISCMI (November 2017), 123–127.
DOI: 10.1109/ISCMI.2017.8279611Google ScholarCross Ref - [17] . 2019. Fusion of spatio-temporal information for Indic word recognition combining online and offline text data. ACM Trans. Asian Low-Reso. 19, 2 (March 2020), 1–24. https://doi.org/10.1145/3364533 Google ScholarDigital Library
- [18] . 2018. Sub-stroke-wise relative feature for online Indic handwriting recognition. ACM Trans. Asian Low-Reso 18, 2 (February 2019), 1–16. https://doi.org/10.1145/3264735 Google ScholarDigital Library
- [19] . 2019. Ship classification in SAR images using a new hybrid CNN–MLP classifier. J. Indian Soc. Remote Sens. 47, (April 2019), 551–562. https://doi.org/10.1007/s12524-018-0891-yGoogle ScholarCross Ref
- [20] . 2018. Supervised and unsupervised learning of fetal heart rate tracings with deep Gaussian processes. Neural Networks Appl., NEUREL (December 2018), 1–6. 10.1109/NEUREL.2018.8586992Google Scholar
- [21] . 2018. Robust deep Gaussian descriptor for texture recognition. Lect. Notes Comput. Sci. 11164, (September 2018), 448–457. https://doi.org/10.1007/978-3-030-00776-8_41Google ScholarDigital Library
- [22] . 2016. Human action recognition with deep action kernel Gaussian process. ICARM - IEEE Int. Conf. Adv. Robot. Mechatronics (August 2016), 165–170. https://doi.org/10.1109/ICARM.2016.7606913Google Scholar
- [23] . 2015. Music emotion recognition using deep Gaussian process. Asia-Pacific Signal Inf. Process. Assoc. Annu. Summit Conf., APSIPA ASC - Proc. (February 2015), 495–498. https://doi.org/10.1109/APSIPA.2015.7415321Google ScholarCross Ref
- [24] . 2017. Educational and non-educational text classification based on deep Gaussian processes. Commun. Comput. Info. Sci. (October 2017), 415–423. https://doi.org/10.1007/978-3-319-70087-8_44Google ScholarDigital Library
- [25] . 2018. Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss. IEEE Trans. Med. Imaging 37, (April 2018), 1348–1357. https://doi.org/10.1109/TMI.2018.2827462Google ScholarCross Ref
- [26] . 2021. Improving skip-gram embeddings using BERT. IEEE ACM Trans. Audio Speech Lang. Process, (March 2021) 29, 1318–1328. https://doi.org/10.1109/TASLP.2021.3065201Google ScholarDigital Library
- [27] . 2020. Pre-trained models for natural language processing: A survey. Sci. China Technol. Sci. (October 2020), 63 (10), 1872–1897. https://doi.org/10.1007/s11431-020-1647-3Google ScholarCross Ref
- [28] . 2020. Transfer learning on Balinese character recognition of lontarmanuscript using MobileNet. ICICoS - Proceeding: Int. Conf. Informatics Comput. Sci. (November 2020). 1--5. https://doi.org/10.1109/ICICoS51170.2020.9299030Google Scholar
- [29] . 2019. Transfer learning using CNN for handwritten Devanagari character recognition. IEEE Int. Conf. Adv. Inf. Technol., ICAIT - Proc., (July 2019), 293–296. https://doi.org/10.1109/ICAIT47043.2019.8987286Google Scholar
- [30] . 2018. Investigation on deep learning for off-line handwritten Arabic character recognition. Cogn. Sys. Res., 50 (August 2018), 180–195. https://doi.org/10.1016/j.cogsys.2017.11.002Google ScholarCross Ref
- [31] . 2017. A novel convolutional neural network architecture for image super-resolution based on channels combination. Int. Conf. Inf. Fusion, Fusion - Proc., (August 2017), 1–8. https://doi.org/10.23919/ICIF.2017.8009771Google ScholarCross Ref
- [32] , et al. 2015. Going deeper with convolutions. Proc. IEEE Comput. Soc. Conf. Comput. Vision Pattern Recognit. (October 2015), 1--9. https://doi.org/10.1109/CVPR.2015.7298594Google ScholarCross Ref
- [33] . 2009. Variational learning of inducing variables in sparse Gaussian processes. Aistats - Int. Conf. Artif. Intell. Stat. (June 2009), 567–574. http://dx.doi.org/Google Scholar
- [34] . 2015. MCMC for variationally sparse Gaussian processes. Proc. Internat. Conf. Intell. Sens. Inf. Process. 1, (June 2015), 1648–1656. https://arxiv.org/abs/1506.04000 Google ScholarDigital Library
- [35] . 2013. Deep Gaussian processes. Proc. Int. Conf. Artificial. Intelligence. Statist (March 2013), 207–215.Google Scholar
- [36] . 2018. Deep Gaussian processes for geophysical parameter retrieval. Dig Int Geosci Remote Sens Symp (IGARSS) (July 2018), 6175–6178. https://doi.org/10.1109/IGARSS.2018.8517647Google Scholar
- [37] . 2012. Stochastic variational inference. Computer Science 14, 1 (May 2012), 1303–1347. https://doi.org/10.1049/iet-cta.2012.0551Google Scholar
- [38] . 2017. AutoGP: Exploring the capabilities and limitations of Gaussian process models. 08, (May 2017). https://arxiv.org/abs/161005392KGoogle Scholar
- [39] . 2018. Deep Gaussian processes with decoupled inducing inputs. (January 2018). https://arxiv.org/abs/1801.02939Google Scholar
- [40] . 2016. Learning convolutional neural networks for graphs. 48, (June 2016), 2014–2023. https://arxiv.org/abs/1605.05273Google Scholar
- [41] 2020. GhostNet: More features from cheap operations. CVPR (August 2020), 1577–1586. https://doi.org/10.1109/CVPR42600.2020.00165Google Scholar
- [42] . 1996. Gaussian processes for regression. In Advances in Neural Information Processing Systems (November 1996), 514–520. Google ScholarDigital Library
- [43] . 2018. Deep Gaussian processes with convolutional kernels, (June 2018). https://arxiv.org/abs/1806.01655Google Scholar
- [44] 2016. Multi-loss regularized deep neural network. IEEE Trans. Circuits Syst. Video Technol. 26, 12 (September 2016), 2273–2283. https://doi.org/10.1109/TCSVT.2015.2477937 Google ScholarDigital Library
- [45] . 2014. Explaining and harnessing adversarial examples. arXiv preprint arXiv (December 2014), 1412.6572. https://arxiv.org/abs/1412.6572Google Scholar
- [46] . 2015. BinaryConnect: Training deep neural networks with binary weights during propagations. Lect. Notes Comput. Sci. 2, (December 2015), 3123–3131. https://arxiv.org/abs/1511.00363Google Scholar
- [47] . 2015. PCANet: A simple deep learning baseline for image classification. IEEE Trans. Image Process. 24, 12 (December 2015), 5017–5032, https://doi.org/10.1109/TIP.2015.2475625Google ScholarDigital Library
- [48] 2016. Multi-loss regularized deep neural network. IEEE Trans. Circuits Syst. Video Technol. 26, 12 (September 2016), 2273–2283. https://doi.org/10.1109/TCSVT.2015.2477937 Google ScholarDigital Library
- [49] . 2017. Inception recurrent convolutional neural network for object recognition, (December 2017). arXiv:1712.09888, https://doi.org/10.1007/s00138-020-01157-3Google Scholar
Index Terms
- Handwritten New Tai Lue Character Recognition Using Convolutional Prior Features and Deep Variationally Sparse Gaussian Process Modeling
Recommendations
Offline handwritten Tai Le character recognition using wavelet deep convolution features and ensemble deep variationally sparse Gaussian processes
AbstractHandwriting recognition is an important application in pattern recognition. Because handwritten Tai Le has similar characters and the proportion of characters with similar shapes is relatively high, this paper proposes a wavelet deep convolution (...
Offline handwritten Tai Le character recognition using ensemble deep learning
AbstractHandwriting recognition is an important area in pattern recognition. For many years, Tai Le has been widely used in Southwest China and Southeast Asia, which makes it of great interest for recognition research. The characteristics of the highly ...
UrduDeepNet: offline handwritten Urdu character recognition using deep neural network
AbstractHandwritten Urdu character recognition system faces several challenges including the writer-dependent variations and non-availability of benchmark databases for cursive writing scripts. In this study, we propose a handwritten Urdu character ...
Comments