This simple project used Support Vector Machine to recognize Sindhi letters. There are three main files. SVM_Trainer.py is used to generate pre-trained model.pkl. SVM_Classifier is used to classify which letter is in the image. model.pkl was generated by SVM_Trainer.py, used by SVM_Classifier and consist file pre-trained on images in img/ folder.
- Python
- OpenCV
- Scikit-Learn
To train the model you need to have .jpg files named 1, 2, 3 etc. and .docx files named at the same way, that consist letters which need to be recognized. This command generate model.pkl:
python SVM_Trainer.py -p user/image/ -q 150
To classify the images you need to run the following command:
python SVM_Classifier.py -i image.jpg
To classify images via Docker container:
docker-compose up