Guddu's profile - activity

I want to develop an end-to-end text recognition (detection and recognition) in natural scene images. Would any body suggest me a simple methodology? I have read so many research papers but my methodology is still not clear.

Training Steps:

1. Input Image
2. Pre-proccessing
3. Text Proposal Generation
4. Feature Extraction
5. Classification
6. Trained Model

Testing Steps:

1. Input Image
2. Pre-proccessing
3. Text Proposal Generation
4. Feature Extraction
5. Classification
6. Recognized Text

Are these steps correct or I am doing any mistake? Any help please? thanks