Data mining for “Non digitally native documents”

*Contact us for further details.

Our software-based methodology enables the transcription of manuscripts, files or old documents difficult to read using standard OCR techniques.

Our proprietary method includes:

🧠 Step 1 – Machine Learning Adapted to the Text We train our models using data from your own collection. PrIx learns to recognize text through its own semantic andgraphical models.
🔍 Step 2 – Direct Search on the Image The system geolocates the zones in the image where a givensearch term is most likely to appear.
📊 Step 3 – Results with Context and Metadata PrIx not only finds the word—it helps structure the information. It can extract names, dates, and relationships, and generate metadata automatically.