An innovative tool that converts images containing numerical digits into MNIST-like formatted data for machine learning applications. An innovative tool that converts images containing numerical digits into MNIST-like formatted data for machine learning applications. This project bridges the gap between real-world digit images and the standardized format of the MNIST dataset, facilitating easier integration and application of machine learning models trained on MNIST.
my project aims to transform any image featuring a number into a format similar to that of the MNIST dataset. This involves detecting the numbers in the image, then scaling and processing them to match the dataset used for model training.
MNIST-Data: Contains sample images from the MNIST dataset for reference and testing.
RawDataForTesting: Includes raw images used to test the algorithm's effectiveness in transforming real-world images into MNIST-compatible format.
src: Contains the core project files including the main script and Jupyter notebooks.
models: Stores trained models, including mnist_model2.h5 which is the output from RunTheModel.py.
notebooks: Jupyter notebooks for detailed exploration and explanation of the project's methodology.
For any queries or collaboration, feel free to contact me:
- Email: [OWD.1@outlook.com]