This page runs the full IDCAPTURE pipeline on every load: crop the ID-number field out of a
captured document photo (icpp_image_crop), OCR just that region
(icpp_ocr_extract, wraps a real Tesseract 5.4.0 install), then insert a record into
ISQL with the extracted text plus the full original photo uploaded into a storage field -
composing three separately-verified building blocks (crop, OCR, the ISQL bridge) rather than one new
native SQL command, since this is naturally a multi-step script-level operation.
OCR of bisected field: ", icpp_ocr_extract("C:\\ai\\claude\\iwebserver\\sample_images\\id_card_cropped.png"), "
"); iprint("| id | id_text (from OCR) |
|---|---|
| ", icpp_sql_get($q,"id"), " | ", icpp_sql_get($q,"id_text"), " |