File size: 17,113 Bytes
9ebbfe5 1d81489 9ebbfe5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 | {
"@context": {
"@language": "en",
"@vocab": "https://schema.org/",
"arrayShape": "cr:arrayShape",
"citeAs": "cr:citeAs",
"column": "cr:column",
"conformsTo": "dct:conformsTo",
"containedIn": "cr:containedIn",
"cr": "http://mlcommons.org/croissant/",
"data": {
"@id": "cr:data",
"@type": "@json"
},
"dataBiases": "cr:dataBiases",
"dataCollection": "cr:dataCollection",
"dataType": {
"@id": "cr:dataType",
"@type": "@vocab"
},
"dct": "http://purl.org/dc/terms/",
"extract": "cr:extract",
"field": "cr:field",
"fileProperty": "cr:fileProperty",
"fileObject": "cr:fileObject",
"fileSet": "cr:fileSet",
"format": "cr:format",
"includes": "cr:includes",
"isArray": "cr:isArray",
"isLiveDataset": "cr:isLiveDataset",
"jsonPath": "cr:jsonPath",
"key": "cr:key",
"md5": "cr:md5",
"parentField": "cr:parentField",
"path": "cr:path",
"personalSensitiveInformation": "cr:personalSensitiveInformation",
"recordSet": "cr:recordSet",
"references": "cr:references",
"regex": "cr:regex",
"repeated": "cr:repeated",
"replace": "cr:replace",
"sc": "https://schema.org/",
"separator": "cr:separator",
"source": "cr:source",
"subField": "cr:subField",
"transform": "cr:transform",
"rai": "http://mlcommons.org/croissant/RAI/",
"prov": "http://www.w3.org/ns/prov#",
"samplingRate": "cr:samplingRate",
"equivalentProperty": "cr:equivalentProperty",
"examples": {
"@id": "cr:examples",
"@type": "@json"
}
},
"@type": "sc:Dataset",
"distribution": [
{
"@type": "cr:FileObject",
"@id": "repo",
"name": "repo",
"description": "The Hugging Face git repository.",
"contentUrl": "https://huggingface.co/datasets/jeffbbrown2/ConnectomeBench2/tree/refs%2Fconvert%2Fparquet",
"encodingFormat": "git+https",
"sha256": "https://github.com/mlcommons/croissant/issues/80"
},
{
"@type": "cr:FileSet",
"@id": "parquet-files-for-config-default",
"containedIn": {
"@id": "repo"
},
"encodingFormat": "application/x-parquet",
"includes": "default/*/*.parquet"
}
],
"recordSet": [
{
"@type": "cr:RecordSet",
"dataType": "cr:Split",
"key": {
"@id": "default_splits/split_name"
},
"@id": "default_splits",
"name": "default_splits",
"description": "Splits for the default config.",
"field": [
{
"@type": "cr:Field",
"@id": "default_splits/split_name",
"dataType": "sc:Text"
}
],
"data": [
{
"default_splits/split_name": "train"
},
{
"default_splits/split_name": "validation"
},
{
"default_splits/split_name": "test"
}
]
},
{
"@type": "cr:RecordSet",
"@id": "default",
"description": "jeffbbrown2/ConnectomeBench2 - 'default' subset\n\nAdditional information:\n- 3 splits: train, validation, test",
"field": [
{
"@type": "cr:Field",
"@id": "default/split",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"fileProperty": "fullpath"
},
"transform": {
"regex": "default/(?:partial-)?(train|validation|test)/.+parquet$"
}
},
"references": {
"field": {
"@id": "default_splits/split_name"
}
}
},
{
"@type": "cr:Field",
"@id": "default/combined_sample_hash",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "combined_sample_hash"
}
}
},
{
"@type": "cr:Field",
"@id": "default/source_archive_sample_hash",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "source_archive_sample_hash"
}
}
},
{
"@type": "cr:Field",
"@id": "default/source_archive",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "source_archive"
}
}
},
{
"@type": "cr:Field",
"@id": "default/sample_type",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "sample_type"
}
}
},
{
"@type": "cr:Field",
"@id": "default/same_neuron",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "same_neuron"
}
}
},
{
"@type": "cr:Field",
"@id": "default/has_single_mask",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "has_single_mask"
}
}
},
{
"@type": "cr:Field",
"@id": "default/has_dual_mask",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "has_dual_mask"
}
}
},
{
"@type": "cr:Field",
"@id": "default/task_routing",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "task_routing"
}
},
"isArray": true,
"arrayShape": "-1"
},
{
"@type": "cr:Field",
"@id": "default/false_split_correction_label",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "false_split_correction_label"
}
}
},
{
"@type": "cr:Field",
"@id": "default/false_merge_identification_label",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "false_merge_identification_label"
}
}
},
{
"@type": "cr:Field",
"@id": "default/split",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "split"
}
}
},
{
"@type": "cr:Field",
"@id": "default/species",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "species"
}
}
},
{
"@type": "cr:Field",
"@id": "default/has_em",
"dataType": "sc:Boolean",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "has_em"
}
}
},
{
"@type": "cr:Field",
"@id": "default/present_slots",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "present_slots"
}
},
"isArray": true,
"arrayShape": "-1"
},
{
"@type": "cr:Field",
"@id": "default/metadata",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "metadata"
}
}
},
{
"@type": "cr:Field",
"@id": "default/geometry",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "geometry"
}
}
},
{
"@type": "cr:Field",
"@id": "default/geometry_single",
"dataType": "sc:Text",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "geometry_single"
}
}
},
{
"@type": "cr:Field",
"@id": "default/em_xy",
"dataType": "sc:ImageObject",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "em_xy"
},
"transform": {
"jsonPath": "bytes"
}
}
},
{
"@type": "cr:Field",
"@id": "default/em_xz",
"dataType": "sc:ImageObject",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "em_xz"
},
"transform": {
"jsonPath": "bytes"
}
}
},
{
"@type": "cr:Field",
"@id": "default/em_yz",
"dataType": "sc:ImageObject",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "em_yz"
},
"transform": {
"jsonPath": "bytes"
}
}
},
{
"@type": "cr:Field",
"@id": "default/em_best",
"dataType": "sc:ImageObject",
"source": {
"fileSet": {
"@id": "parquet-files-for-config-default"
},
"extract": {
"column": "em_best"
},
"transform": {
"jsonPath": "bytes"
}
}
}
]
}
],
"conformsTo": "http://mlcommons.org/croissant/1.1",
"name": "ConnectomeBench2",
"description": "\n\t\n\t\t\n\t\tConnectomeBench2\n\t\n\nConnectomeBench2 is a unified benchmark for automated proofreading of connectomic neural-segmentation data. 401,170 samples across 4 species (mouse, fly, human, zebrafish) and 5 sample types (real merge edits, real split edits, synthetic adjacent / junction / synapse controls), with the associated mesh geometry and electron-microscopy (EM) renderings.\nDownstream trainers should treat this dataset as the single source of truth for sample identity, labels\u2026 See the full description on the dataset page: https://huggingface.co/datasets/jeffbbrown2/ConnectomeBench2.",
"alternateName": [
"jeffbbrown2/ConnectomeBench2",
"ConnectomeBench2"
],
"creator": {
"@type": "Person",
"name": "Jeff Brown",
"url": "https://huggingface.co/jeffbbrown2"
},
"keywords": [
"other",
"100K - 1M",
"parquet",
"Image",
"Text",
"3D",
"Datasets",
"Dask",
"Polars",
"Croissant",
"\ud83c\uddfa\ud83c\uddf8 Region: US",
"connectomics",
"proofreading",
"3d",
"electron-microscopy",
"mesh"
],
"license": "https://choosealicense.com/licenses/other/",
"url": "https://huggingface.co/datasets/jeffbbrown2/ConnectomeBench2",
"citeAs": "Brown, J., Farkas, T., Razgar, G., Boyden, E. S. ConnectomeBench2: A unified benchmark for automated connectomic proofreading. (2026, in submission). Brown J. and Farkas T. contributed equally as first authors.",
"datePublished": "2026-05-07",
"version": "1.0.0",
"rai:dataLimitations": "Sample distribution is skewed across species. Fly (45%) and mouse (37.5%) dominate, zebrafish (12.4%) and human (5.1%) are underrepresented. EM views are present on only 63% of samples; tasks that require EM should filter via `has_em` (controls = 100%, real edits = 23-38%). Real edits reflect human proofreader judgment and may carry their selection biases. Junction and synapse controls are synthetic and meant as task-specific negatives, not as a faithful sample of the natural false-merge / false-junction error distribution. Not recommended for: clinical or diagnostic use; tasks requiring representativeness of any specific neural circuit.",
"rai:dataBiases": "(1) Species imbalance: fly (45%) and mouse (37.5%) dominate; zebrafish (12.4%) and human (5.1%) are sparse, so models trained on the full mix may perform much worse on zebrafish/human. (2) Brain-region bias: each upstream connectome covers a specific brain region (mouse visual cortex, fly central brain, H01 cortex, zebrafish larval brain). Thus, the data is not representative of \"all neural tissue\". (3) Edit selection bias: real human proofreading edits reflect what proofreaders chose to fix, which is likely biased toward visually obvious or task-relevant segments and away from edits that are hard to spot. (4) Modality bias: EM is present on all synthetic controls but only ~30% of real edits, so EM-trained models see a different sample mix than geometry-only training; this can confound comparisons.",
"rai:personalSensitiveInformation": "No personally identifiable information. Source data is mesh geometry and electron-microscopy of brain tissue from non-human animals (fly, mouse, zebrafish) and one previously-published, de-identified human cortical biopsy sample (H01, Harvard/Lichtman lab). No demographic, clinical, or behavioral data is included.",
"rai:dataUseCases": "The captured construct is human proofreader decisions on real connectome volumes: real merge_edit/split_edit rows are actual curator-applied corrections; control rows follow documented synthetic recipes (see prov:wasGeneratedBy). The dataset validly trains/evaluates models for imitating proofreader decisions on volumes similar to the upstream sources. It does not capture the broader \"all segmentation errors\" construct, as real edits skew toward errors humans noticed, and controls reflect specific construction recipes, not natural error distributions. Not validated for: detecting errors human proofreaders missed; stand-alone proofreading deployment; neuron-type classification; any clinical use.",
"rai:dataSocialImpact": "Positive: accelerates connectomics research by reducing the manual proofreading bottleneck and standardizing evaluation of automated methods. Risks are minimal as data is neural mesh geometry, not subjects. Main indirect risk is over-reliance on automation reducing human review quality. Human-in-the-loop deployment is recommended.",
"rai:hasSyntheticData": true,
"prov:wasDerivedFrom": [
"https://www.microns-explorer.org/cortical-mm3",
"https://flywire.ai/",
"https://h01-release.storage.googleapis.com/landing.html",
"https://fish1-release.storage.googleapis.com/index.html"
],
"prov:wasGeneratedBy": "Real edits (merge_edit, split_edit) come from CAVE proofreading edit history of the upstream connectomes. Adjacent controls are sampled from segments adjacent to real edits (synthetic negatives for merge correction). Junction controls are sampled from putative junctions in proofread neurons (synthetic negatives for false-merge ID). Synapse controls are pre/post-synaptic root pairs where presynaptic and postsynaptic neurons differ (synthetic negatives for merge correction). All samples are rendered into: (a) 3 orthographic views (front/side/top) of the mesh as a 7-channel float16 array (silhouette + depth + xyz normals + per-segment masks), stored as compressed npz; (b) 4 EM slices (xy, xz, yz, oblique 'best') as 3-channel uint8 PNGs (raw EM + per-segment masks). Train/val/test split is by 50\u00b5m spatial cubes randomly assigned to splits to prevent leakage from spatial proximity. Pipeline scripts are available at https://github.com/timfarkas/ConnectomeBench2/tree/main/scripts/neurips/."
} |