What Is Document Detection?
Document Detection is a Filestack intelligence task that finds a document in a photo, straightens it so it fills the image, and cleans it up so that OCR reads the text more accurately.
What the Task Does
- Finds the edges of the document in the photo
- Corrects the perspective so the document fills the output image
- Reduces noise and distortion (preprocessing)
- Returns the coordinates of the document when
coords:trueis set
How It Runs
Synchronously in the Processing API with doc_detection, or as a task in Filestack Workflows. The Processing API request needs a security policy and signature. Defaults: coords:false, preprocess:true.
Image Size Limit
Input images can be at most 2000×2000 pixels. Chain the resize task before doc_detection to bring a larger image within the limit.
Use Cases
- Preparing phone photos of documents for OCR
- Digitizing receipts, forms and IDs
- Cropping scanned pages to the document edges
Related terms
Optical Character Recognition (OCR)
Intelligence service extracting printed and handwritten text from images.
Intelligence (AI) Service
Machine learning tasks that tag, caption, moderate, scan and read files.
Workflows
Background processing of uploaded files with tasks and logic rules.
More in Intelligence
This page is one of 94 terms in the Filestack glossary.
Browse all terms