Authors: 1st Author- Shafaque Ahmareen, 2nd Author – Alreem Alabdouli (me), 3rd Author – Sirisha Polturi
Paper in the 5th International Conference on Image Processing and Capsule Networks Titled "MNSIT Handwritten Digit Recognition using CNN"
Start Date: May 2024
End Date: 3 – 4 July 2024
Research Field: CNN, Deep Learning
Subfield: Computer Science
Status: Published
Publisher: IEEE
Publication Date: September 2024
DOI: 10.1109/ICIPCN63822.2024.00018
URL : IEE Xplore-MNSIT Handwritten Digit Recognition using CNN
Research question: This research study uses the MNIST dataset to investigate how Convolutional Neural Networks (CNN) might be used to recognize handwritten numbers. From collecting data to evaluating models, it covers it all. We feed the preprocessed MNIST dataset into a CNN.
Paper Summary
This research uses the MNIST dataset to implement Convolutional Neural Networks (CNN) for recognizing handwritten digits. The project involves fine-tuning CNN layers, such as convolutional, max pooling, and dense layers, using various hyperparameters like optimizers and learning rates to enhance model performance. Batch normalization techniques were also explored to increase accuracy and generalization. The study demonstrates how CNN models can effectively classify handwritten digits, achieving high accuracy rates on both training and test datasets, laying the groundwork for future improvements in image processing and real-time applications.