cnrocr

Container number detection and recognition — ISO 6346 end-to-end, ONNX only.

A D-FINE detector locates the number, ISO type code, owner code and serial on the container; a CRNN recognizer reads each crop with a spec-constrained CTC beam search. Fragments split across panels are merged back into a single number and validated against the ISO 6346 check digit.

No PyTorch required. The runtime is onnxruntime + numpy + pillow.

Install

pip install cnrocr           # CPU
pip install cnrocr[gpu]      # NVIDIA CUDA via onnxruntime-gpu

Model weights (~113 MB) are not bundled in the wheel. They are downloaded on first use and cached locally:

cnrocr models download       # optional — happens automatically otherwise

Python API

from cnrocr import ContainerOCR

ocr = ContainerOCR()                      # weights resolved from cache
result = ocr.read("gate_cam.jpg")

for c in result.containers:
    print(c.number, c.iso_type, c.confidence, c.needs_review)
# TGHU8913884 22G1 0.9997 False

Multiple images

results = ocr.read_many(["a.jpg", "b.jpg", "c.jpg"], batch_size=8)

read_many treats the images as unrelated — N images in, N results out.

Multi-view fusion

When several cameras photograph the same container, fuse them into one answer instead of voting on strings:

mv = ocr.read_multiview(["cam1.jpg", "cam2.jpg", "cam3.jpg"])
print(mv.number, mv.agreement, mv.mode)

Beam-search candidates from every view are summed in log space, so a character that one view is unsure about can be settled by the others. If the views appear to be looking at different containers, they are not fused — mv.consensus becomes False rather than producing a confident wrong answer.

Review triage

A check digit alone is not enough. The constrained decoder only emits spec-conforming candidates, so when the true string is absent from the beam it will confidently output a plausible wrong number that still passes the check digit. needs_review combines the check digit, the spec flag and a confidence floor:

if c.needs_review:
    print(c.review_reason)     # "신뢰도 낮음 (0.612 < 0.7)"

Owner code registry (optional)

Real-world owner codes are registered with the BIC. Supplying the list filters out invented codes that would otherwise pass both the format check and the check digit:

from cnrocr import OwnerCodeRegistry
ocr = ContainerOCR(registry=OwnerCodeRegistry.from_file("bic_codes.txt"))

The list is not shipped with this package.

Command line

cnrocr read gate_cam.jpg
cnrocr read *.jpg --json --device cuda
cnrocr multiview cam1.jpg cam2.jpg cam3.jpg
cnrocr models status
cnrocr check                    # diagnose install, providers, cache

--fail-on-review makes the process exit with code 2 when any result needs human review, which is convenient in batch pipelines.

Evaluation limit

Without a licence, cnrocr processes 10 images per day. The counter is per image, not per call, and resets at 00:00 UTC.

cnrocr license      # licence status and today's usage

When the quota runs out the process exits with code 3 and prints how to ask for a licence. Nothing is processed on a call that would exceed the quota — it is all or nothing, so a refused call costs no quota.

To request an unrestricted licence, email vislab2026@gmail.com with your name, organisation and intended use. You will receive a key:

# Windows
setx CNROCR_LICENSE "eyJlbWFpbCI6..."

# macOS / Linux
export CNROCR_LICENSE='eyJlbWFpbCI6...'

The key may also be saved to a file named license in the cache directory (cnrocr models path shows where). Verify with cnrocr check.

Weights and caching

Platform	Location
Cache (Windows)	`%LOCALAPPDATA%\cnrocr\Cache\models\<set>`
Cache (macOS)	`~/Library/Caches/cnrocr/models/<set>`
Cache (Linux)	`~/.cache/cnrocr/models/<set>`

Every file is verified against a SHA-256 recorded in the wheel. Released assets are immutable: a new model set ships under a new tag and a new library version, so upgrading never invalidates an existing install.

The weights are encrypted and are decrypted into memory when a session is built. The cache holds ciphertext only; no plaintext model is written to disk.

Environment overrides:

Variable	Effect
`CNROCR_MODEL_DIR`	Use this directory as-is; never download
`CNROCR_CACHE_DIR`	Relocate the cache root
`CNROCR_WEIGHTS_BASE_URL`	Fetch weights from somewhere else (`file://` works)

License

Proprietary. Evaluation and non-commercial research use only — see LICENSE. Contact the copyright holder for commercial licensing.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

cnrocr-0.3.2-cp313-cp313-win_amd64.whl (681.9 kB view details)

Uploaded Aug 7, 2026 CPython 3.13Windows x86-64

cnrocr-0.3.2-cp313-cp313-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (5.4 MB view details)

Uploaded Aug 7, 2026 CPython 3.13manylinux: glibc 2.17+ x86-64manylinux: glibc 2.28+ x86-64

cnrocr-0.3.2-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl (5.3 MB view details)

Uploaded Aug 7, 2026 CPython 3.13manylinux: glibc 2.17+ ARM64manylinux: glibc 2.28+ ARM64

cnrocr-0.3.2-cp313-cp313-macosx_11_0_arm64.whl (721.9 kB view details)

Uploaded Aug 7, 2026 CPython 3.13macOS 11.0+ ARM64

cnrocr-0.3.2-cp312-cp312-win_amd64.whl (687.2 kB view details)

Uploaded Aug 7, 2026 CPython 3.12Windows x86-64

cnrocr-0.3.2-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (5.5 MB view details)

Uploaded Aug 7, 2026 CPython 3.12manylinux: glibc 2.17+ x86-64manylinux: glibc 2.28+ x86-64

cnrocr-0.3.2-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl (5.4 MB view details)

Uploaded Aug 7, 2026 CPython 3.12manylinux: glibc 2.17+ ARM64manylinux: glibc 2.28+ ARM64

cnrocr-0.3.2-cp312-cp312-macosx_11_0_arm64.whl (733.4 kB view details)

Uploaded Aug 7, 2026 CPython 3.12macOS 11.0+ ARM64

cnrocr-0.3.2-cp311-cp311-win_amd64.whl (702.3 kB view details)

Uploaded Aug 7, 2026 CPython 3.11Windows x86-64

cnrocr-0.3.2-cp311-cp311-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (5.4 MB view details)

Uploaded Aug 7, 2026 CPython 3.11manylinux: glibc 2.17+ x86-64manylinux: glibc 2.28+ x86-64

cnrocr-0.3.2-cp311-cp311-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl (5.4 MB view details)

Uploaded Aug 7, 2026 CPython 3.11manylinux: glibc 2.17+ ARM64manylinux: glibc 2.28+ ARM64

cnrocr-0.3.2-cp311-cp311-macosx_11_0_arm64.whl (742.2 kB view details)

Uploaded Aug 7, 2026 CPython 3.11macOS 11.0+ ARM64

cnrocr-0.3.2-cp310-cp310-win_amd64.whl (701.8 kB view details)

Uploaded Aug 7, 2026 CPython 3.10Windows x86-64

cnrocr-0.3.2-cp310-cp310-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (5.1 MB view details)

Uploaded Aug 7, 2026 CPython 3.10manylinux: glibc 2.17+ x86-64manylinux: glibc 2.28+ x86-64

cnrocr-0.3.2-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl (5.1 MB view details)

Uploaded Aug 7, 2026 CPython 3.10manylinux: glibc 2.17+ ARM64manylinux: glibc 2.28+ ARM64

cnrocr-0.3.2-cp310-cp310-macosx_11_0_arm64.whl (748.0 kB view details)

Uploaded Aug 7, 2026 CPython 3.10macOS 11.0+ ARM64

File details

Details for the file cnrocr-0.3.2-cp313-cp313-win_amd64.whl.

File metadata

Download URL: cnrocr-0.3.2-cp313-cp313-win_amd64.whl
Upload date: Aug 7, 2026
Size: 681.9 kB
Tags: CPython 3.13, Windows x86-64
Uploaded using Trusted Publishing? No
Uploaded via: twine/7.0.0 CPython/3.11.13

File hashes

Hashes for cnrocr-0.3.2-cp313-cp313-win_amd64.whl
Algorithm	Hash digest
SHA256	`be689efbdc3ee70584dc35f19d89160b01cdafe5e65dda98177e99594d7f8178`
MD5	`245050301c26f3d822bd648736b461ef`
BLAKE2b-256	`40f5e01a73a5a3f6f3153eda8905b967feee2710e97247f4fd6262e7c66cecd0`

Algorithm	Hash digest
SHA256	`4620e69e9e48dbcc69c4d457c8fe50d1f30d1d7b29a772f3432dc23e5b6888a7`
MD5	`6ca818d83d0b8ad591418dd2f470fb85`
BLAKE2b-256	`0017f0d62b9e7f2fbcc40d47b0de85065f5f59a13c2bbce3816853fc3fc88e28`

Algorithm	Hash digest
SHA256	`2c66631c9ec99e9893a61c5de3a7a601efc3a123b0aaafa4bbc583c5cc9ee3b9`
MD5	`1da955f9d129b2a62b88c71f32c30efa`
BLAKE2b-256	`5c3d3962690365f6d3eb1082ea462fd8296cd1e062c5bf333bd9b6b9229d79ea`

Algorithm	Hash digest
SHA256	`d3a08bc5f9e54e1ab15596fc682c9b31e2f1cfc3bfc5b7eebbcfbe9e1cd925f0`
MD5	`0271070ebe2b61bfd67d32a5076f7e15`
BLAKE2b-256	`b8487227631b077318d1b7c195f3b6476ec426c1cfe1530e5be7e450b4b4dae8`

Algorithm	Hash digest
SHA256	`6feb9e8c5a87ee20053a4ed5eb7ba96841a34af2e7040f7bbfa8cd351472dbde`
MD5	`335487af1bf455d104b4daafef95aaa5`
BLAKE2b-256	`2d7d721f964141634f466f871b1fe65536b4d7efd30faa9e35fb455963ccc57a`

Algorithm	Hash digest
SHA256	`effaa79c7d2f938b8d496b37f9803fe506c83773674e0dfef9b461136eef0fb8`
MD5	`78e4694f91858d5038e9aedb32b4814e`
BLAKE2b-256	`eb623ade808872b4fb833d3e21d4e845d17b240378e30c1f53b40a000f976ca5`

Algorithm	Hash digest
SHA256	`91f76d07ff8e40785b8c23968031536812c34b54bde651a5b6946070712f7a8c`
MD5	`8331ceb767ca8652cadb66d0add76a4f`
BLAKE2b-256	`3eb5833f5b2cd711ff3ab97b0cc5418fac59f9e44b4da3abccf9103fa47b8d07`

Algorithm	Hash digest
SHA256	`94d9a038f73af86d2b5131edb082dc56471ff7ab6e52e9bf5b58311a2238e5b5`
MD5	`46eefd064885e8c4b25c7d19e3a02b7f`
BLAKE2b-256	`3db69ec2112161a8df602ef62432afcbcd9692270c08a8025c1a903f970ccde2`

Algorithm	Hash digest
SHA256	`c5a22650ab16fe6a03e2acb598c5554bc2a800d4fad4cae6c424ac532285e623`
MD5	`9ad834f24b2607176410088b4c2c28ad`
BLAKE2b-256	`09cd2b2a319663c51f80ad878050a9187302ac8488441b6299f36add901fc2d5`

Algorithm	Hash digest
SHA256	`347681641e6e05ed644501540fa2ce7c8be3d2c2e46668fd7e7d1c5c59c39cc3`
MD5	`36e487e46edceb5769fa1aceda54c871`
BLAKE2b-256	`f2f25c99a1e8b40296eb0d1578e4285b562b64ab09d618a55ecd0083be035193`

Algorithm	Hash digest
SHA256	`5a2a23918726f79b0b91526610d2b5c393d534c231fd321df55ff868b1539f95`
MD5	`56d60942ee2f67ee950e61126fbd5995`
BLAKE2b-256	`0d97a7b59d908c5dbada1c3163013bde7b7a66716fd096210f418e5187a0a938`

Algorithm	Hash digest
SHA256	`a641d66b7bb109be8de0d7337ff8ced7662b0371a9804576c26c2399a8c23b3e`
MD5	`a6e33c0064b80d10d765309f8c3a9ace`
BLAKE2b-256	`fec270186c3ae6a1b553729a3d2952cb122e0a0ca2b2fa26cd24ecfcff749bb1`

Algorithm	Hash digest
SHA256	`c0de6ea02d0e14ca9bb11bc322f5a513319c9ce8f0c635bb8c07454d13ee617b`
MD5	`a1a009b7b84761d7c7b650a787900fe8`
BLAKE2b-256	`a7e51b1c71194d1cc6fa6e853b0686c247eae4b9a27c168b5d2e257b134897c4`

Algorithm	Hash digest
SHA256	`adea756398f9b210d6d5c7bc822a76378b04c09634c10513b4fbd4b811fd6c21`
MD5	`b6af717c7fca1f682831efc3d2abeeaa`
BLAKE2b-256	`5ee917dbc7ffe9cca300b916f93941643f68f03dc86631182880dbb6158ba3b3`

Algorithm	Hash digest
SHA256	`baa0497d53f23d99af353c5716625b7770890159cd69593f5c75258ce081f6e1`
MD5	`e6d36845d91bf24b8fbd87784f97094b`
BLAKE2b-256	`45144e182c7c2acfc9fecec0b3b01fef6feac12e764720807ee0f43ec53bd30d`

Algorithm	Hash digest
SHA256	`e926bdd98d0dae7c3267a1013a917b7ab5f8fe34b65d58320b18a07d165b7084`
MD5	`a804948e4662d211f1b3a7ee9c5d6721`
BLAKE2b-256	`e5902cd599b270c65fe4568f43134e9c58087890c58d159e9fc83a70da8a1b66`

cnrocr 0.3.2

cnrocr

Install

Python API

Multiple images

Multi-view fusion

Review triage

Owner code registry (optional)

Command line

Evaluation limit

Weights and caching

License

Project links

Key dates

1 maintainer

Credits

License expression

Requires

Provides Extra

Tags

Classifiers

Download files

Source Distributions

Built Distributions

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

Release history Release notifications | RSS feed