json3 KB

receipt (gaussian blur) ground truth

The exact answer for the blur variant of this receipt: every one of the 23 words with its string and pixel box, plus the transform that produced this variant. Authored before the image existed, so it is ground truth by construction rather than a second opinion.

OCR testing

json

application/json

3 KB

Document: receipt
Variant: blur
Word Count: 23
Font: consola.ttf
Box Format: axis-aligned [x0,y0,x1,y1] in pixels
Source Plate: nss-invoice-desk_00001_.png

Binary json: no in-browser preview. Download it above to open in a compatible application.

Specifications

Document: receipt
Variant: blur
Word Count: 23
Font: consola.ttf
Box Format: axis-aligned [x0,y0,x1,y1] in pixels
Source Plate: nss-invoice-desk_00001_.png
Synthetic: true
Disclosure: Synthetic. Background is AI-generated; all text was typeset programmatically.
Schema Version: 1
Transform Gaussian Blur Radius Px: 1.8

Testing contract

Reference control

Scenario: Parse the file and check that every box lies inside the companion image and that no two boxes are identical.
Expected result: It lists 23 words, each with its string and an axis-aligned pixel box, for the blur variant; the recorded transform is gaussian blur radius px 1.8. The geometry was authored BEFORE the image was drawn, so it is ground truth by construction rather than a second recogniser's opinion.

What is a .json file?

JSON (JavaScript Object Notation) is a lightweight, text-based data-interchange format representing objects, arrays, strings, numbers, booleans, and null. It is language-independent, human-readable, and the dominant format for web APIs and configuration. It requires a single well-formed root value.

How to use this file

Use an example JSON file to test parsers and serializers, schema validation, Unicode and number-precision handling, and API request or response processing.

How to use this file for testing

“receipt (gaussian blur) ground truth” is a deterministic Novus Examples fixture for OCR testing. Image-only 'scanned' documents paired with their text source, for measuring OCR accuracy against a known ground truth.

Documented properties for this file: JSON · 3,123 bytes. Compare results against paired or grouped companions on this page when present (clean↔damaged, searchable↔scanned, or format twins) so scores stay reproducible across runs.

Download the file once, keep the path stable in CI or local scripts, and treat the spec table as the contract: dimensions, seeds, field lists, and roles are intentional. Corrupt or invalid samples are labelled as such, expect parsers to fail loudly rather than silently accept them.

Data fixtures document their exact quirks (delimiters, encodings, null handling, schema, and row counts) in the spec table. Point your parser or importer at the file and assert it handles the documented edge cases; clean and deliberately-messy siblings make before/after diffs straightforward.

Run OCR on the scanned or image twin and score the output against the searchable or text ground truth on this page; the documented rotation, grain, and text content are the reference.

Code examples

import json

with open("ocr-receipt-blur.json") as f:
    data = json.load(f)
print(type(data), len(data))

Generated by generation/ocr_fixtures_wave.py. Free for any use, no attribution required, license.