HMC-Grad is a tool designed to automate the grading process of handwritten multiple-choice tests. It uses a Convolutional Neural Network (CNN) trained in PyTorch on the EMNIST dataset and OpenCV for image processing. The tool comes with a Gradio interface for ease of use.
-
No Specialized Hardware: HMC-Grad eliminates the requirement for specialized hardware like OMR scanners, allowing automation through common devices such as smartphones.
-
Simplified Answer Sheet Requirements: Removing the need for custom printed answer sheets, HMC-Grad processes answers on standard yellow notepads.
-
Automated Grading: Input a correction key (as a string of text) and the images (JPG) of the answer sheets, and HMC-Grad will provide the correctness and score for each one, as well as grading annotations.
-
Item and Score Analysis: Obtain detailed item-wise analysis along with an overall score analysis.
-
Export Results: HMC-Grad generates CSV and XLSX files containing the results, making it easy to integrate with other tools or platforms.
- Google Colab environment
-
Open the provided notebook (
hmc_grad.ipynb) in Google Colaboratory. -
Execute all the cells in the notebook to run HMC-Grad in a user-friendly Gradio interface.
-
Scroll down and open the generated Gradio public URL.
-
Upload your correction key and the images of the answer sheets into the respective input fields. The system can handle any number of image inputs.
-
Download the results you need.
-
Open the hugging face space link.
-
Upload the correction key and the answer sheet images.
-
Download the results.
This project is licensed under the MIT License - see the LICENSE file for details.
For any inquiries or issues, please contact gabriel.edradan05@gmail.com.