| --- |
| license: other |
| license_name: mixed-annotations-cc-by-4.0 |
| license_link: LICENSE |
| pretty_name: Broken QR Codes (104) — a hard real-world QR benchmark |
| task_categories: |
| - image-to-text |
| - object-detection |
| tags: |
| - qr-code |
| - barcode |
| - benchmark |
| - document-understanding |
| - computer-vision |
| - robustness |
| size_categories: |
| - n<1K |
| configs: |
| - config_name: default |
| data_files: |
| - split: test |
| path: data/** |
| --- |
| |
| # Broken QR Codes (104) |
|
|
| **104 real QR code images that ordinary decoders fail on**, with byte-level ground |
| truth where a public reference decoder could establish it, geometric annotations, and |
| per-image results for four reference decoders. |
|
|
| This is a **test set, not a training set**. Every image here was collected *because* |
| something failed on it, so the distribution is deliberately adversarial rather than |
| representative. |
|
|
| ## Why this is hard |
|
|
| **78 of the 104 images are read by none of the four reference decoders**: Apple |
| Vision, Apple CIDetector, and WeChat's CNN and classical detectors. The best of them, |
| Apple Vision, reads 21. |
|
|
| The failure modes are the interesting part: artistic QR codes where the module grid is |
| deliberately obscured by generated imagery, physical damage, extreme perspective, low |
| contrast, motion blur, partial occlusion, screen glare, and resolutions from 0.05 to 16 |
| megapixels. |
|
|
| ## What's in each row |
|
|
| | field | notes | |
| |---|---| |
| | `image` | the picture itself | |
| | `file_name`, `image_id`, `width`, `height` | | |
| | `has_payload`, `payload`, `payload_bytes_hex` | **26 of 104** have a payload: the images at least one reference decoder reads | |
| | `qr_version` | QR version, on **19** images | |
| | `quad` | the symbol's four corners in pixels, on **22** images | |
| | `finder_centers` | the three finder-pattern centres, on **19** images | |
| | `u_px_per_module` | pixels per module, on **19** images | |
| | `quad_source` | the reference decoder whose reading gave the geometry | |
| | `solved_by_*` | four booleans, one per reference decoder | |
| | `n_decoders_solving` | 0–4 | |
|
|
| ### Only 26 images can be checked |
|
|
| A payload is given only where a reference decoder reads the code, so **26 is the |
| ceiling for scoring**. A decoder that reads one of the other 78 images cannot be |
| credited from this dataset, because nothing here says what the right content is. |
| Report the denominator you used. |
|
|
| `payload` and `payload_bytes_hex` are not equally populated: 26 rows carry a payload |
| string and 20 carry the raw bytes. Where both exist they agree; where only the string |
| exists, binary payloads may not round-trip through UTF-8, so **compare bytes when you |
| have them**. |
|
|
| ## Reference decoder results |
|
|
| Measured on one machine (Apple M2, macOS 14.3.1), decode call only, image already |
| resident in memory. |
|
|
| | decoder | solves / 104 | false | p50 µs | |
| |---|---:|---:|---:| |
| | Apple Vision | 21 | 0 | 11,195 | |
| | Apple CIDetector | 18 | 0 | 7,168 | |
| | WeChat (CNN) | 14 | 0 | — | |
| | WeChat (classical) | 13 | 0 | — | |
|
|
| Apple returns a string plus a raw codeword stream and is scored on both surfaces; |
| WeChat is string-only. |
|
|
| ## Usage |
|
|
| ```python |
| from datasets import load_dataset |
| |
| ds = load_dataset("devmandan/broken-qr-codes-104", split="test") |
| |
| # the checkable subset: the only fair denominator |
| checkable = ds.filter(lambda r: r["has_payload"]) # 26 |
| |
| # the images every reference decoder failed |
| unsolved = ds.filter(lambda r: r["n_decoders_solving"] == 0) # 78 |
| |
| # geometry-supervised subset |
| geo = ds.filter(lambda r: r["quad"] is not None) # 22 |
| ``` |
|
|
| ## Provenance, licensing, and takedown |
|
|
| **Please read this before redistributing.** |
|
|
| The **annotations** in this dataset — payloads, quads, finder centres, versions, |
| module scales, decoder results — are original work and are offered under |
| **CC BY 4.0**. |
|
|
| The **images** are a different matter and are not uniformly owned by the |
| publisher of this dataset. They were collected from public web sources as |
| examples of QR codes that fail to decode, including artistic/generated QR codes |
| whose individual authorship is not documented. They are published here for |
| **research and benchmarking**, which is the purpose for which they were |
| gathered. No claim of ownership is made over them. |
|
|
| If you own an image in this set and want it removed, open a discussion on this |
| repository and it will be taken down. |
|
|
| **Payload contents are real.** Several decoded payloads are live third-party URLs |
| that were encoded in the original codes, some carrying opaque identifiers. They |
| are reproduced verbatim because a QR benchmark whose payloads have been rewritten |
| cannot be used to check a decoder. Do not treat any URL here as endorsed, safe, or |
| maintained, and be careful about resolving them automatically. |
|
|
| ## Citation |
|
|
| ```bibtex |
| @misc{broken_qr_codes_104, |
| title = {Broken QR Codes (104): a hard real-world QR decoding benchmark}, |
| author = {Legut, Daniel}, |
| year = {2026}, |
| url = {https://huggingface.co/datasets/devmandan/broken-qr-codes-104} |
| } |
| ``` |
|
|
