Graphology is the study of human personality through handwriting. It is considered a projective personality test that helps reveal various traits based on writing style and patterns.
This repository is an effort to apply Machine Learning to analyze handwriting using established rules from graphology. The system extracts characters from handwritten text, processes them using OCR, and applies trained models to interpret personality traits.
📄 A more detailed explanation of the project can be found in the project report.
- Uses graphological principles to interpret individual characters
- Implements Photo OCR (TesseractJS) to extract handwritten symbols
- Applies Neural Networks for classification and personality inference
- Developed a simple web-based platform for real-time testing
- Explores both CNN and shallow NN architectures, with real-time symbol classification
| Category | Stack / Tools |
|---|---|
| Machine Learning | Keras, TensorFlow, Neural Networks, CNN |
| Image Processing | TesseractJS (OCR), Canvas API |
| Dataset | EMNIST (Letters) |
| Web Platform | HTML, CSS, JavaScript, TensorFlowJS |
- Expand to analyze multiple letters for deeper personality insight
- Improve model accuracy with richer datasets and augmentation
- Integrate Google OCR for higher symbol recognition confidence
- Refine UI/UX of the web app for wider usability
- Develop per-letter model pipelines to support full-text diagnosis