tesseract don't detect words in my tif

[Jorge Infante]

Hi. 

I'm trying to use tesseract to find alphanumeric information in a .tif. I find that if I reduce the size .tif, find as many words using the full .tif. The idea is that I am inspecting the label of a plane, looking for information. Reviewing the full tagline, I find much less information than if I use a near info area that interests me. I can improve this somehow? The reality is that I need to inspect the full tagline, as it has a fixed position in the plane, unlike internal tables that may be changing position. So my technique is to find words (with a hocr output, which gives me the bounding box) give me information location of those boxes.

Thanks a lot

PD: If I check with tiffinfo .tif not differ the properties of the small picture about the big picture. Is there any pattern tif quality? I can not include, for now, the plane, because it is private taxpayer information.