Published on: April 2026
HANDWRITTEN DIGIT RECOGNITION USING NEURAL NETWORKS
K. Nikitha M. Saniya M. Sandeep A. Naveen
Y. V. S. Durga Prasad
Article Status
Available Documents
Abstract
To overcome these challenges, this project presents a neural network-based solution for recognizing handwritten digits using supervised learning. The system begins by preprocessing input images through normalization, reshaping, and label encoding to ensure effective model training. A feedforward neural network is then developed using TensorFlow/Keras, incorporating multiple hidden layers with activation functions that enable the model to learn complex relationships within the data. The model is trained using an efficient optimization technique and evaluated over several training cycles to improve its predictive performance.
How to Cite this Paper
Nikitha, K., Saniya, M., Sandeep, M. & Naveen, A. (2026). Handwritten Digit Recognition Using Neural Networks. International Journal of Creative and Open Research in Engineering and Management, <i>02</i>(04). https://doi.org/10.55041/ijcope.v2i4.201
Nikitha, K., et al.. "Handwritten Digit Recognition Using Neural Networks." International Journal of Creative and Open Research in Engineering and Management, vol. 02, no. 04, 2026, pp. . doi:https://doi.org/10.55041/ijcope.v2i4.201.
Nikitha, K.,M. Saniya,M. Sandeep, and A. Naveen. "Handwritten Digit Recognition Using Neural Networks." International Journal of Creative and Open Research in Engineering and Management 02, no. 04 (2026). https://doi.org/https://doi.org/10.55041/ijcope.v2i4.201.
Search & Index
References
- Zhang, C., et al. (2026). ResNet18-ThunderSVM: Hybrid Intelligence for Handwritten Digit Recognition. Technology: ResNet18 + ThunderSVM. Limitation: High computational complexity and sensitivity to handwriting variations.
- Parveen, H., et al. (2026). CNN-Based Approach for Handwritten Digit Recognition. Technology: CNN with convolutional and pooling layers. Limitation: Restricted to clean MNIST dataset, struggles with noisy inputs.
- Joel, T. O., et al. (2025). Performance Comparison of CNN and LSTM for Handwritten Digit Classification. Technology: CNN and LSTM. Limitation: LSTM misclassifies static images and has slower processing.
- Kochkorova, A., & Toumpa, A. (2025). Data Augmentation for Handwritten Digit Recognition. Technology: CNN with augmentation (rotation, scaling, shifting). Limitation: Some augmentations introduce artifacts; limited exploration of hybrid models.
- Ben Noureddine, D. (2025). Comparative Analysis of ML, CNN, Vision Transformers and Hybrid Models. Technology: ML, CNN, ViT, Hybrid Models. Limitation: Traditional ML weak at feature extraction; ensemble approaches add complexity.
- Sindhu, D. G., & Manjunatha, G. C. (2025). Handwritten Digit Recognition Using Multilayer Neural Networks with Keras and TensorFlow. Technology: MLP with Keras/TensorFlow. Limitation: Overfitting risk, struggles with similar-looking digits, heavy preprocessing dependency.
- Desale, A., et al. (2024). Hybrid CNN + LSTM Model for Hindi Handwritten Digit Recognition. Technology: Hybrid CNN + LSTM. Limitation: Limited to Hindi digits, high computational resource requirement.
- Sharma, R., et al. (2024). Machine Learning and Deep Learning-Based Handwritten Digit Recognition System. Technology: Random Forest, SVM, CNN. Limitation: Classical models struggle with noisy data and generalization.
- Kottakota, S., et al. (2023). NeuroWrite: Deep Neural Network Combining CNNs and RNNs for Predictive Handwritten Digit Classification. Technology: CNN + RNN. Limitation: Requires extensive training data and computational power.
- Arora, P., et al. (2023). Ensemble-Based Handwritten Digit Recognition System Using CNN and SVM.
Ethical Compliance & Review Process
- •All submissions are screened under plagiarism detection.
- •Review follows editorial policy.
- •Authors retain copyright.
- •Peer Review Type: Double-Blind Peer Review
- •Published on: Apr 10 2026
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. You are free to share and adapt this work for non-commercial purposes with proper attribution.