Influencing results by giving suggestions

[goncalo-oliveira]

Depending on the quality of the photo and the steadiness of the photographer, the text recognized can sometimes be misinterpreted. For example, an I might be read as an L or a 1.

If it was possible to suggest that the text we are scanning is all capitals, for example, we could potentially minimize this.

If there's any way of doing this, I probably missed, so any hints are welcome.