We need the quality estimation to be part of the README

[MansMeg]

Now we only have speaker mappings of the quality dimensions. We should include the others as well:

• [ ] OCR quality
• [ ] segmentation quality
• [ ] segment classification accuracy