an AI / ML Inference Platform

About

Employ convenient cloud-based inference services to solve challenging problems.

  • Optical Character Recognition (OCR)

  • Handwritten Text Recognition (HTR)

  • Template Parsing

doc-parser

Parses fixed-format images (e.g., ID cards, passports) and extracts the text from various fields of interest via OCR by relying on a known, standardized template.

E.g.: a Romanian ID card (left) can be parsed into the following JSON (right):

{
    "match_score": 0.5011664535634441,
    "last_name_field": "VASILESCU",
    "first_name_field": "ELENA",
    "cnp_field": "2410415400342",
    
    "MRZ": 
    {
        "doctype": "ID",
        "country": "ROU",
        "last_name": "VASILESCU",
        "first_name": "ELENA",
        "series": "SS",
        "number": "099994",
        "nationality": "ROU",
        "birthday": "410415",
        "sex": "F",
        "validity": "770415",
        "cnp_first_digit": "2",
        "cnp_suffix": "400342",
        "series_number_crc_match": 0,
        "birthday_crc_match": 0,
        "validity_crc_match": 0,
        "final_crc_match": 1
    }
}

?

requests served... so far

Let’s talk

If you find overfitted.io to be matching your business requirements, we’d love to hear from you!