Improved Model Configuration Strategies for Kannada Handwritten Numeral Recognition


  • Gopal Dadarao Upadhye Pimpri Chinchwad College of Engineering, Pune, Maharashtra, 411044, India
  • Uday V. Kulkarni Shri Guru Gobind Singhji Institute of Engineering and Technology (SGGSIET), Nanded, India
  • Deepak T. Mane JSPM's Rajarshi Shahu College of Engineering, Pune, India



Numeral recognition, particle swarm optimization, convolutional autoencoder, Kannada numerals


Handwritten numeral recognition has been an important area in the domain of pattern classification. The task becomes even more daunting when working with non-Roman numerals. While convolutional neural networks are the preferred choice for modeling the image data, the conception of techniques to obtain faster convergence and accurate results still poses an enigma to the researchers. In this paper, we present new methods for the initialization and the optimization of the traditional convolutional neural network architecture to obtain better results for Kannada numeral images. Specifically, we propose two different methods- an encoderdecoder setup for unsupervised training and weight initialization, and a particle swarm optimization strategy for choosing the ideal architecture configuration of the CNN. Unsupervised initial training of the architecture helps for a faster convergence owing to more task-suited weights as compared to random initialization while the optimization strategy is helpful to reduce the time required for the manual iterative approach of architecture selection. The proposed setup is trained on varying handwritten Kannada numerals. The proposed approaches are evaluated on two different datasets: a standard Dig-MNIST dataset and a custom-built dataset. Significant improvements across multiple performance metrics are observed in our proposed system over the traditional CNN training setup. The improvement in results makes a strong case for relying on such methods for faster and more accurate training and inference of digit classification, especially when working in the absence of transfer learning.

Author Biographies

Gopal Dadarao Upadhye, Pimpri Chinchwad College of Engineering, Pune, Maharashtra, 411044, India

Computer Engineering and Assistant Professor

Uday V. Kulkarni, Shri Guru Gobind Singhji Institute of Engineering and Technology (SGGSIET), Nanded, India

Computer Science & Engineering and Professor

Deepak T. Mane, JSPM's Rajarshi Shahu College of Engineering, Pune, India

Computer Engineering and Associate Professor


Chen M, Shi X, Zhang Y, Wu D, Guizani M (2017).

Deep features learning for medical image analysis

with convolutional autoencoder neural network.

IEEE Transactions on Big Data.

D. T. Mane UVK (2018). Pattern recognition of iris

flower using neural network based particle swarm

optimization. International Journal of Computer

Sciences and Engineering 6.

Dhandra B, Mukarambi G, Hangarge M (2011).

Zone based features for handwritten and printed

mixed kannada digits recognition. In: IJCA

Proceedings on International Conference on VLSI,

Communications and Instrumentation (ICVCI),

no. 7.

Eberhart R, Kennedy J (1995). A new optimizer using

particle swarm theory. In: MHS’95. Proceedings

of the Sixth International Symposium on Micro

Machine and Human Science. Ieee.

El-Sawy A, Loey M, El-Bakry H (2017). Arabic

handwritten characters recognition using

convolutional neural network. WSEAS

Transactions on Computer Research 5:11–9.

Erhan D, Courville A, Bengio Y, Vincent P (2010).

Why does unsupervised pre-training help deep

learning? In: Proceedings of the Thirteenth

International Conference on Artificial Intelligence

and Statistics.

Ganesh A, Jadhav AR, Pragadeesh KC (2016). Deep

learning approach for recognition of handwritten

kannada numerals. In: International Conference on

Soft Computing and Pattern Recognition. Springer.

Gurudath K, Ravi D (2016). Isolated digits recognition

in kannada language. International Journal of

Computer Applications 140.

Hallur VC, Hegadi R (2013). Kannada handwritten

digits recognition: neural network approach. Int J

Sci Res 417 419 .

Hallur VC, Hegadi R (2014). Offline kannada

handwritten numeral recognition: holistic

approach. In: Proceeding of Second International

Conference on Emerging Research in Computing,

Information, Communication and Applications,

vol. 3.

Karthik S, Murthy KS (2015). Handwritten kannada

numerals recognition using histogram of oriented

gradient descriptors and support vector machines.

In: Emerging ICT for Bridging the FutureProceedings of the 49th Annual Convention of

the Computer Society of India CSI Volume 2.


Kavya T, Pratibha V, Priyadarshini B, Vijaya Bharathi

M, Vijayalakshmi G (2016). Kannada characters

and numerical recognition system using hybrid

zone-wise feature extraction and fused classifier.

Int J Eng Res Technol 5.

Kennedy J, Eberhart R (1995). Particle swarm

optimization. In: Proceedings of ICNN’95-

International Conference on Neural Networks,

vol. 4. IEEE.

Killedar S, Deshapande S (2015). Kannada

handwritten numerals recognition and translation

using template matching. International Journal on

Recent Technologies in Mechanical and Electrical

Engineering 2:77–80.

Mamatha H, Srirangaprasad S, Srikantamurthy K

(2013). Data fusion based framework for

the recognition of isolated handwritten kannada

numerals. Int J Adv Comput Sci Appl 4:174–82.

Mane D, Kulkarni UV (2020). A survey on

supervised convolutional neural network and its

major applications. In: Deep Learning and Neural

Networks: Concepts, Methodologies, Tools, and

Applications. IGI Global, 1058–71.

Mukarambi G, Dhandrab V, Hangarge M (2011).

Recognition system for handwritten and printed

kannada numerals and vowels. Int J Mach Intell


Nair V, Hinton GE (2010). Rectified linear units

improve restricted boltzmann machines. In: ICML.

Prabhu VU (2019). Kannada-mnist: A new

handwritten digits dataset for the kannada

language. arXiv preprint arXiv190801242 .

Prasanna Kumar K (2013). Algorithm to identify

kannada vowels using minimum features

extraction method .

Ragha LR, Sasikumar M (2010). Adapting moments

for handwritten kannada kagunita recognition. In:

Second International Conference on Machine

Learning and Computing. IEEE.

Rajput G, Horakeri R, Chandrakant S (2010). Printed

and handwritten kannada numeral recognition

using crack codes and fourier descriptors plate.

International Journal of Computer Application

IJCA on Recent Trends in Image Processing and

Pattern Recognition RTIPPR :53–8.

Ratadiya P, Asawa K, Nikhal O (2020). A

decentralized aggregation mechanism for training

deep learning models using smart contract system

for bank loan prediction. arXiv preprint

arXiv201110981 .

Ratadiya P, Mishra D (2019). An attention ensemble

based approach for multilabel profanity detection.

In: 2019 International Conference on Data Mining

Workshops (ICDMW). IEEE.

Roy A, Dutta D, Choudhury K (2013). Training

artificial neural network using particle swarm

optimization algorithm. International Journal of

Advanced Research in Computer Science and

Software Engineering 3.

Sheshadri K, Ambekar PKT, Prasad DP, Kumar RP

(2010). An ocr system for printed kannada using

k-means clustering. In: 2010 IEEE International

Conference on Industrial Technology. IEEE.

Shettar S, Basavaprasad B, Bhagya H (2015).

Recognition of printed kannada numerals by

nearest neighbor method. In: Proceedings of

the International Conference on Computational

Systems for Health Sustainability.

Srivastava N, Hinton G, Krizhevsky A, Sutskever I,

Salakhutdinov R (2014). Dropout: a simple way

to prevent neural networks from overfitting. The

journal of machine learning research 15:1929–58.




How to Cite

Upadhye, G. D., Kulkarni, U. V., & Mane, D. T. (2021). Improved Model Configuration Strategies for Kannada Handwritten Numeral Recognition. Image Analysis and Stereology, 40(3), 181–191.



Original Research Paper