Research and reproducibility

Datasets

The data behind OTS-HOMR: scanned and rendered staff images, their musical transcriptions, and structured notation labels. Download a built corpus to train or evaluate a model, or use its source repository to reconstruct the data.

Staff-crop downloads

Each window contains an image, recognition tokens and a structured-notation sidecar. These releases use homr.notation-sidecar.v6. Train and validation scores are kept separate. The GitHub repositories contain the build and verification tools; Cloudflare R2 hosts the compressed data.

Scanned staff crops

OpenScore String Quartets

ossq-scanned-v2

Scanned string-quartet staves paired with symbolic transcriptions. Version 2 restores stem-direction labels while retaining the original crop images and recognition tokens.

Train crops
35,226
Validation
3,911
Compressed
887.8 MB

76 training scores · 9 validation scores

Ordered indexes: train · validation

Download archives (6 shards)
OpenScore String Quartets archives and SHA-256 checksums
Archive / SHA-256SplitSize
train-0000.tar.zstbc809df59e87c13f0130918b656e8f4e851c22f202a53091ee5ee50a66a431f1Train168.7 MB
train-0001.tar.zst6f9ac54a319a1764c5c1fb7238ad380f0b9f415108b2a2a4528f1e0e23e18ce2Train159.5 MB
train-0002.tar.zstb2f3b0d1e2dec76cae4f607b195b297c576318bd36f4c3c26d801b1f89819742Train165.5 MB
train-0003.tar.zst5e2f9a94007a4ec4b22b316a5832cb3fc9230f930220dc54584c0d2a335c1c64Train170.0 MB
train-0004.tar.zst26b74b535513d179fe41d6dd6112c7e6e119ab79bd3a3343b0d6c61eb3bf8be1Train138.7 MB
valid-0000.tar.zst68519b83959c122481687386a976080518eecf3b5cf1b90f9e3535fcae68b284Validation85.4 MB

Scanned staff crops

OpenScore Lieder

lieder-pd212-v8

Voice-and-piano staff crops from the 212 IMSLP sources included in the public v8 release, paired with OpenScore Lieder transcriptions.

Train crops
3,567
Validation
300
Compressed
481.9 MB

199 training scores · 13 validation scores

Ordered indexes: train · validation

Download archives (3 shards)
OpenScore Lieder archives and SHA-256 checksums
Archive / SHA-256SplitSize
train-0000.tar.zst96697d92f155cf46ccdaf1d91c2521d46eb06f386dac7d3ffc6c9470d1c725c2Train377.1 MB
train-0001.tar.zstcf6b2d13fc34c0a89951025716f4b5a7f48d1cf1aa74882663595fc58e9c57ecTrain71.7 MB
valid-0000.tar.zsta8b53148b7e409e1f3112aa9dfb51db656c0b4fbc94a4c41dd15e5af89329495Validation33.0 MB

Rendered staff crops

PDMX · v1

pdmx-homr-v1

Verovio engravings of public-domain and CC0 MusicXML, cut into overlapping eight-measure windows. This is the PDMX version used in the published staff-recognition benchmark.

Train crops
41,635
Validation
3,422
Compressed
1.94 GB

6,376 training scores · 632 validation scores

Ordered indexes: train · validation

Download archives (8 shards)
PDMX · v1 archives and SHA-256 checksums
Archive / SHA-256SplitSize
train-0000.tar.zst2657e4d189b014bcd9f5f3c42151350ed2ddb12d0d15d74364d87b25934bebbdTrain279.3 MB
train-0001.tar.zsta8accb2c87dd54d23d04df9a6196c52067be2d642338423c2f7730893260a5f4Train275.2 MB
train-0002.tar.zstfd9a11fe55e56b67bb848f50bcc1d8d6114ba948739a2938942eb0b0914018e2Train276.2 MB
train-0003.tar.zstc16be6a015649fe3e068ffdfe251e02268e3b765da62b71a17c3d66dd4c7b722Train278.2 MB
train-0004.tar.zst7b02040bbc974c5c3fe24ab2d4640fcb9ae7d06d2669ec0dbc6669cf5c0d1ec9Train275.9 MB
train-0005.tar.zst76d883b1e45d2118dd946b4af3272d251b29ae82a7b6c41c8e4ac43f3bfaab5cTrain274.9 MB
train-0006.tar.zst9a40669b0c706e55dcf2728e5ffad97ee7ddfab1136fd66513cec9c791c51ce1Train142.4 MB
valid-0000.tar.zst5922d6eb4c6cd5ed74affd3f66d862cd68e16e8618e0f0f7f051c98785533c9dValidation142.4 MB

Rendered staff crops

PDMX · expanded v2

pdmx-homr-v2

An expanded release with up to six tracks and more source scores. Its validation slice has different membership from v1; choose the version named by the recipe or benchmark you are reproducing.

Train crops
565,522
Validation
1,829
Compressed
19.07 GB

31,342 training scores · 184 validation scores

Download archives (118 shards)
PDMX · expanded v2 archives and SHA-256 checksums
Archive / SHA-256SplitSize
train-0000.tar.zst438e1b85738531b8a57c219ebd8e1f6542dfe6ed3879d3466c164e2d81e36bf3Train164.0 MB
train-0001.tar.zst1afe9af78fb08ce4c1aef24e7ed79858d4803aae77eb3f45dcaca50047381c68Train166.4 MB
train-0002.tar.zstb987f6ab2b528fa015aeae6c7ebe82c201fa43372c29cfa0471407acbffcebeaTrain162.8 MB
train-0003.tar.zstf89f14606444a0bc201e970f167eb3a0929fa55dd5b2f1bff0d1479b99fe105aTrain164.5 MB
train-0004.tar.zst9bc6e4399f99e70de0490290c1ee9bca4f846365ce96a2c0e6036dda20f3df9fTrain163.0 MB
train-0005.tar.zstbcd69e0d8a6f07b3f6aaf249093f18a467222f6c9e5371df4f10930229c6d9dcTrain164.4 MB
train-0006.tar.zst689f6b87f810262ebfe60e777bfd907e427535b070c305f8926e2af2f1777568Train165.4 MB
train-0007.tar.zsta38bc12df3f259a26379b038076e2e0653c96b3d98194a804ef1cd5e07df7c1aTrain163.3 MB
train-0008.tar.zst9289cdc64af3e12584ae11c6c1fe7bb2401c9237ae961fae510c116d62a53409Train159.4 MB
train-0009.tar.zst2a44bebd06dbe198cac3950b2f676a517ed53999e6b228b8d8bb2d3c009ee195Train165.7 MB
train-0010.tar.zste8316ee6e43616624f8bb393c0ef222defdbce5cd0237371cc3b01adb01337b2Train166.0 MB
train-0011.tar.zstb94d7bece907958924adf883d72883ea4e7e4f9c4590373e56e30f5a70edc054Train165.9 MB
train-0012.tar.zst2ac9f96b5b650eeea6d365c69de818ed8416dc0a972e6682d2ee12fbb0b5b914Train164.7 MB
train-0013.tar.zst22418bb1dd7dd91f990bb22e84ffcb944fcd3e454e9212df51be58ebd83b49c1Train158.5 MB
train-0014.tar.zstf54da5bfbc690e9ec60572e8f300b52ec36ce924a6c08ee7aeb2d84d1976b9fdTrain161.9 MB
train-0015.tar.zst8f3346c023588f62c59d0b546b6402f353958dab48a689ff96380f3963c4e96aTrain164.6 MB
train-0016.tar.zsted2923460b1e65a979546375e53f7ebfb2d92a37b617b3157b9104aaee82642aTrain164.2 MB
train-0017.tar.zstfa9f646b5cf5e88a458f006ca62ce52d733d18e32bd4e528c5206f92e5b42544Train162.1 MB
train-0018.tar.zst89866d5d6afbc3b629158d2fd3e6ccfd8f0cc0f73dd156320840c73e1a309835Train165.9 MB
train-0019.tar.zstec59525b935224806b143654dd5263196c231e292e3fcc239655d66e70bbfd7eTrain166.8 MB
train-0020.tar.zstbbf5a91af4659f8baa27254c54bb4f6d5f25a3ae7d2461988d77f7a2dcbac794Train165.5 MB
train-0021.tar.zst3892212668e1580645e86d0d0939a4563e3a57f45c7ebe3b03746da28a1f7682Train165.5 MB
train-0022.tar.zst177772c223c4498afc7c689322ccd7333d653781cfcd96a5d0eb4a98ddb9b5bdTrain164.2 MB
train-0023.tar.zstba430867607ae665a31b63cccaeb4d39817221d9b4348c4c01133e1182c5979eTrain163.0 MB
train-0024.tar.zstf27f94b86bc5f5541dbca31463bab4932295fb1bad9ad108a2722fe4e50ab830Train161.8 MB
train-0025.tar.zstd04e4eeca22b2b3fe84c9c1e79b6c9705c94c1bec5adf227f749b89c66eef9deTrain160.7 MB
train-0026.tar.zst6feaf63cd8667567d831c7135746570d2adfad4843b9ba5063883347f692c911Train166.0 MB
train-0027.tar.zstb9486956c3d18ab26956a4abcb06a8701014714e8db9a8882eb668c44bec2909Train163.4 MB
train-0028.tar.zst33e427a2613bdac175997d0afbb2722becee136388eac4d920ded85ace647f75Train162.1 MB
train-0029.tar.zstaa121d51c6c262d6c22159a1a5664fa785bdf7e6c2a9c5f4c69d125d240d9468Train163.3 MB
train-0030.tar.zst299ab5f9197d479e12dafdbb3230a19c99d2b92ba5411212b82fa6b5dea8c0f2Train164.8 MB
train-0031.tar.zst3ad6b0e33f603d71009cc4fb8fb7d2eb587ec576799d0a78e556cb5be19869a9Train164.5 MB
train-0032.tar.zstcbeb67609478c78b581a9b8174150a73d37fd2d505257dcee5536b08057ad7ceTrain165.1 MB
train-0033.tar.zst6db45e6724b9ba8dacea5db6bc4c4d2a41444fdd83147fb349ed017baa6adeb4Train166.1 MB
train-0034.tar.zstddd7aed38b902df81d97c4cfee312d9957d0e6e2b10d0418af9ce380628bff00Train160.5 MB
train-0035.tar.zst896f21a7b5514cb5f6b026660a34f1f0a39a55edda4f65d3115fb8a9d9cdb5ceTrain161.5 MB
train-0036.tar.zst1fde44730dd4149bc0ae554110d6ee97a99bd5417863eb0c402098a401e36614Train157.1 MB
train-0037.tar.zst1fc5b9df2bb13cfb70cb1174718b148e16681ab41bba26222e2f11896ef6ccc0Train164.2 MB
train-0038.tar.zstc090acee44a444b5e9495adcdc6d3aab6d0a8a570235fe5d3b8b3f59ee3121f5Train164.6 MB
train-0039.tar.zstb6ee4666987eaa3c409255a9c7fed7212dbb344cf275001a55c023c7136633b8Train161.8 MB
train-0040.tar.zst43b0be2180da5f0ff3b7501420a6da79a937eccc77c63d912966ba77c454363bTrain165.7 MB
train-0041.tar.zst1758a7ea27b5b0a7aaefa2bc1ecf6d17e87380c8f3ffbb54833bfdb08c4ae2acTrain161.4 MB
train-0042.tar.zst635f775127afdf195ad3f1e43a6193bc23ca0e364b10725cf2aa15977c5dd4e0Train164.5 MB
train-0043.tar.zst8af8070bab01a4614a81a9eeba0e4a3cdac4416d50350ab8cf7ade272b950331Train163.5 MB
train-0044.tar.zst8488287dba075c87bb5dc64dc54ac5e584ec706f050e2f132bbd4a1721b7a0d5Train163.0 MB
train-0045.tar.zsta00d3d9992969eb71c6c8bce376794a9bd35b151f5024cfba2c21b6e8dbaf341Train160.1 MB
train-0046.tar.zst851144d3c506838e840a09bff2e5e71664c3e833a7093fc9bb1a30117b7fdafeTrain163.1 MB
train-0047.tar.zst68f55639b3cba42f03a1c7c9a0c35f12d9ba8d6981dfd3c0522819d807068eccTrain165.8 MB
train-0048.tar.zstd7c46ec9feb51edc084943679bd10dbba1ecc898bf82b61cefdab9863cf94323Train163.9 MB
train-0049.tar.zst393f43f5530f9d48087421402c7a681b45e9e66e649fc74ccf3a9aee53f5f74eTrain163.5 MB
train-0050.tar.zst527ca44aa33041a076b055c123092be66586611569c9cb694162583b40140930Train162.8 MB
train-0051.tar.zst033f2ff4e4ad56e0029a87397c9562eea7fa69ea5212378e42c4bbc31c297b6fTrain163.2 MB
train-0052.tar.zste452a4966f43d2e35385ac3be3bb1b4ea21919cc71a2dc8bb48c5de7154d75ccTrain162.6 MB
train-0053.tar.zste150523f64c50e7e7bfa75f181441b89596f64a30088bde053df10fbc8f0d118Train165.3 MB
train-0054.tar.zst8db56d3cc492bb5ac08abb8f9351afb4caf64e2d11f02676fbfa925d8ffc6fa9Train163.0 MB
train-0055.tar.zstda173077aec9fa7aea841d4bd8cc6c5e990c146db9282bb9792ae769f5691de6Train165.9 MB
train-0056.tar.zst3ad5f9540a3d45558ece1124ca5a2dc706cb7abdaeecf346656548aac1c2d5b9Train159.0 MB
train-0057.tar.zst7503af530737d0d95d302db3a70fdf86344373d911b1bb79089ac2d22627aeb2Train163.6 MB
train-0058.tar.zstf3e8d202d8e4bd581701e108e83da82e0c01b4da092ba0a4b5915d553a447c6eTrain160.8 MB
train-0059.tar.zst3c87a878b3038e4fb8784fbc1d6fa3418964b43034ccc761f447d797ad81b76cTrain162.5 MB
train-0060.tar.zstb6c87a25bfd2420c4caa610a7a1f609d557d07c5dc04e4770fc6ece2a892c138Train163.4 MB
train-0061.tar.zst6b0245a49ee83ccf83c0b33755eadba939bbba13afaa23b477ccd316d8826f03Train162.4 MB
train-0062.tar.zsta3c15709b2b3f14a03fd43d863d6b7149b60b2dd2d1897aea0405fe536a80786Train167.7 MB
train-0063.tar.zstc2bfc0239196962c0fa00fe28198f6d9901b4895298595d60b1a33aab5104686Train162.1 MB
train-0064.tar.zst155537ad2a2290cd8259715a80101ed730a8215aa27a7552539c667414e1f220Train164.2 MB
train-0065.tar.zsta40a62ad339d1bd22cafc21e1a89223e85c749e8ab60cca0a3c5cd208fc949c7Train159.5 MB
train-0066.tar.zstfbc1b38042e9b3a884d8bd60a233a3e3dc51a0f72d41642e950680a3c926a5fbTrain164.2 MB
train-0067.tar.zst61caedaba29ff431b24bcd989888f3750977bc31d2d64ed6797c0b4181216040Train163.4 MB
train-0068.tar.zstdd1361e55f9bfceb0848705a07dbefa145c8f55fee683cce6faa4e3a28ef153eTrain162.3 MB
train-0069.tar.zst9cc6eaa4dfa36a21117e7e4574233c3acb9fcf4ccbe954e51952c16946c2d116Train163.1 MB
train-0070.tar.zst836bf64a9041ae843e4b59fb47a789b79a04411133b325107e62cbceddde2789Train166.1 MB
train-0071.tar.zst3fa6081c86e6031d0b04fed3efd358c9ed22169511720326de70b9c927ffbc88Train164.3 MB
train-0072.tar.zst8a348893622387ae7b406b7905a2fcad592a8c57e6a7527d00957810492699f1Train164.1 MB
train-0073.tar.zst4c5cb63a31aaa2301503344e38d9d4668a68b8283f224779cb87b2b46a79629eTrain165.0 MB
train-0074.tar.zst8702e68f344f941c607c5bd94f695f8c2772bda8eff3b9f9e2481488c0ee012aTrain166.3 MB
train-0075.tar.zstbebffd0d0267fd03226affa639154cec8845418bdb24555f13bf83c0b0753cf1Train163.1 MB
train-0076.tar.zst04c75628b4a44e79cd1196976ab3899f0e050a6cf30a09afb84b18c069cd9fe4Train161.5 MB
train-0077.tar.zst896c899b0d3912c6d0deb80f55d799d03c654ab52c0f7856e2484396abb8b59fTrain162.7 MB
train-0078.tar.zst15e345dfe327c6517a810d2ca1bf86090a0d8e8151764c5403b7a83cf933f24fTrain166.7 MB
train-0079.tar.zste88c446ccb50636dc72d733cbc4bae6f9d428ec57230914ca7cf4d6f61cc163aTrain164.1 MB
train-0080.tar.zstedb1a5451a46576fa1a9f784636add3453ed6e144317ffd2e9e508d77a280353Train164.9 MB
train-0081.tar.zst924d63fd52166db154862035d7228653beb10908828d947691cbde8d86c9cff1Train163.7 MB
train-0082.tar.zstc2e0b01dc73d4f179148f069e0c5e985aca70f7495dec23f9b072b623681293dTrain165.6 MB
train-0083.tar.zst8d48ff443ab4172d0db64da50c6b416e09f5fe4eeec7d4b0fb3ab54bce9e8ffdTrain163.6 MB
train-0084.tar.zst3d47ce334d9f4e981816585e53bd2b254144475722f61691822d86570dfec90aTrain160.3 MB
train-0085.tar.zst1ea1c5cd5b640bf206d125e6e7349b23b5dc8780a4f9ef4165dfbbe340e5ea06Train161.8 MB
train-0086.tar.zst6c1876d9dd3576b58e943194f2856b9bac65c162fc2f2d4ffd1d64fd128f83e3Train165.9 MB
train-0087.tar.zst952a52f5380a2c96db8e59ba123bc21140011f086c2f7735211d48d3c03fc12cTrain163.7 MB
train-0088.tar.zst44f482c0dc443e94bf29a4bbfcf9ffcf54f477101ad84b1d6d93ef457fc3835aTrain161.9 MB
train-0089.tar.zst72c03f0e5d4dc1fb55d99dfab4eac57123c70c81da03c46db01e1601af9fe3fcTrain162.0 MB
train-0090.tar.zst2268f4b74f7dc892fd0ad5b3c8447551c2665cc96e05bcc3d0b446d6821f1d72Train161.6 MB
train-0091.tar.zst203fff0f58792b194d5bb9bfdcdd2140dba4b680e9457dca4b2760c1c4aa985aTrain164.0 MB
train-0092.tar.zstce7c24157f54b6c645ba7c49e6e873b8d65b7c145f8ac6660eed5439014af445Train164.0 MB
train-0093.tar.zst7b9bee9059c93d0f4f2a9f82d69e7415a0217719f8318db4ed4e43145869bf64Train159.8 MB
train-0094.tar.zst326a83226cd31bba527f3fbd1395216c6cf692e7c8f0ff92e8fccb24f4382d41Train164.0 MB
train-0095.tar.zstefd755e743420cce3650462f92522cfcaf53e2e7483a6b66325f3fa3dcfe774bTrain164.9 MB
train-0096.tar.zstb50e36f579645e32c76016d97929cda2537801ca6850e8f6b2ed221e06c711a1Train163.2 MB
train-0097.tar.zst3fbd5d9a8fd3a2953a7358bcd4a8f45dc19dc338109276754d479f45c58bc2a0Train162.6 MB
train-0098.tar.zstc024da0d134a814394461bd78c21da9cda24421a804bb33530a6a1f9373daecdTrain163.0 MB
train-0099.tar.zsta6385529575a3f1e39e02bbe142e9b0657a6d5f96d0b086acfe063973e725d66Train159.9 MB
train-0100.tar.zst0783301ea4e36030ba58f0b0ffa836c46ac6e919ac2179f20f8a950f5ed740ecTrain160.6 MB
train-0101.tar.zstdbfd97a8902c0ce33093ca758331835e150cb24f9e7e5d654fb87ff8d2f173beTrain164.8 MB
train-0102.tar.zst5a432bfb487f1c51a6eaf8e0f559c77b450e4f7a578115b417f5c4b529c15343Train164.2 MB
train-0103.tar.zstf043d7a866e29d1053809d6f706dd7f4bd6420921cd3d7149c665a9f7937e683Train165.3 MB
train-0104.tar.zstc549b43df2e24fbd7dd1c69b922d47dab052bb8aa9604c610a93c58a5c2e163fTrain163.6 MB
train-0105.tar.zst5cf845da944d698735e76dacff55466b635ca696d0753d1f4b17b779868f682cTrain164.4 MB
train-0106.tar.zsta7e4a8d6384910efc67ba363171932c5cde13f2857b5dea074286015438d7412Train164.2 MB
train-0107.tar.zst2b9b6f382ad1d88222c3f6925dd1ca6874f88b50eea4a1f95119b63bdbb071f5Train164.4 MB
train-0108.tar.zst9115211146df01398b511cf077b691ba52302161d9683d4ecc9263fd25ceb2beTrain166.0 MB
train-0109.tar.zst6943b9c8a26e302c60b7d16376ab9ef03e8aa882e3cb44fcebc993b8845ecb18Train164.6 MB
train-0110.tar.zstfcc60ba17145ffd8929e9f34cbfcd953575ae639667b69d951e33f780d132501Train161.2 MB
train-0111.tar.zst53466064b3a93b181ed2f6972c0fe3ccb44bd31c676c47af958f19f32f532017Train164.5 MB
train-0112.tar.zst864d41f71f5d586c28dc8e7d03ac3bcb0a718e7c719fa0afcbe42570c92c1480Train164.4 MB
train-0113.tar.zstdf512244c5314418d6e12728acdb85c7571d922527a55e6b2475a04b7cd2d30dTrain163.7 MB
train-0114.tar.zst36ed4ffa655dca206dcbc2cbf176e503bcf70765cb40fa3a2fb8fcd331c5c361Train165.2 MB
train-0115.tar.zst45a9ef83c4edc2f4be45672a8fc2bb90f1dbbb3c056a145d0cd525c78e33d9a5Train164.5 MB
train-0116.tar.zstf66f73e0aff306d1eccb04ba8fa2c3118a588088000fdfa9bb7ea94fe04fb07fTrain22.4 MB
valid-0000.tar.zsta1f524b711bcaa0aadf291bc98b8d2e7eccb84fda2010e0a800cb6caafafade3Validation78.1 MB

Hugging Face archives

These are separately versioned archives and source records. Use their dataset cards to select the right release; the historical OSSQ archive and the R2 v2 corpus have different training-row counts.

Lieder source provenance

Source-file records, rights-audit ledgers and reconstruction indexes for all 215 historical IMSLP inputs. This repository contains provenance files, rather than scan PDFs or staff-crop images; use the Lieder R2 release for the built crops.

Browse this archive version

GrandStaff LMX

Grand-staff images with LMX, MusicXML and kern representations, including distorted images. Training, development and test batches are recorded in the dataset card.

Browse this archive version

Download the OSSQ v1 archive (794 MB) · Unpacked-tree manifest

Reproduce the data

  1. Choose a version. Use the release named by your training recipe. Keep its source lock, indexes and sidecar schema together.
  2. Fetch and verify. Clone the build repository, download the corpus shards and check their SHA-256 digests against the repository’s lock.
  3. Preserve the splits. The indexes keep entire scores in one split. Keep that separation when sampling crops or comparing checkpoints.
  4. Rebuild when needed. Each repository documents its source inputs, pinned converter and rendering environment. OSSQ and Lieder use scanned material; PDMX uses Verovio 6.2.1 engravings.
git clone https://github.com/OurTextScores/ossq-omr-data.git
cd ossq-omr-data
make fetch-corpus
make verify-corpus

# To fetch only the validation split:
make fetch-corpus SPLIT=valid

The same fetch commands are available in the Lieder and PDMX build repositories. PDMX’s default lock fetches v1; the expanded v2 release has its own manifest and build recipe above.

The dataset cards and source manifests record attribution and rights. Build-code licences and source-image licences are separate; the Lieder source ledger records individual IMSLP files, while PDMX records the selected scores’ public-domain and CC0 provenance.

The OTS-HOMR scanner recipe

OTS-HOMR v11 combines a scan-adapted HOMR recognition core, structured notation heads, and a separate lyric, direction and dynamics review pipeline. Its minimal source recipe records the runtime, dependencies, calibration and exact model identities used for inference.

Minimal HOMR repository · v11 recipe · Source manifest · Model revisions and checksums

git clone --branch release/v11-minimal https://github.com/OurTextScores/homr.git
cd homr
python release/verify_source.py
python release/build_source.py

This recipe reproduces the scanner source and its pinned-weight inference setup. Rebuilding a training corpus is a separate step from retraining the historical v11 weights. The OurTextScores application provides the scanner queue, review interface and score export; its integration is described in the recipe.

OurTextScores application source · Recognition benchmark · Evaluation method