ocr op• Data kinds: matrix × table → table
• Call: import fullseye as fs; fs.ledger.ocr_classify(features, model, k=1, blur=None) -> 'dict' (to call the implementation directly, import ocrclassic; ocrclassic.ocr_classify(features, model, k=1, blur=None) -> 'dict'; from the registry, opssymbol.get("ocr_classify"))
• GPU: none (CPU reference implementation only)
Classify feature vectors by k nearest neighbours and reject by distance and distance ratio (table).
> The detailed description below is the original text — the summary and the headings are translated.
features : (d,) か (m, d)、d = FEATURE_LENGTH。
model : ocr_train の返り値。
k : 近傍数(奇数推奨)。多数決、同票は最近傍のクラス。
blur : 既知(または推定済み)のぼけ σ/字高。与えると、学習のぼけ格子でそれと隣り合う
格子点までの試料だけを照合に使う(ocr_read_line は行全体で推定して渡す)。
Restrict templates to the blur band around blur (sigma / char height).
dict
labels (list of str、棄却は "?"), raw_labels(棄却前の判定),
distance(判定クラスの最近傍距離。model["weights"] で重みづけした対角マハラノビス距離),
margin(他クラス最近傍距離 / 判定クラス最近傍距離、1 に近いほど境界上), rejected。
棄却条件: distance > reject_distance_by_class[判定字] または margin < min_margin_ratio。
• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).
• Operator provenance and references — the sources of the research/methods this op family came from.
• The canonical algorithm (author, year) and its uses are named in the family usage guide above.
• poc_symbol_lot_ocr — py -3.11 examples/poc_symbol_lot_ocr.py
table as input)render_symbol · symbol_read_trials · symbol_read_pod · ocr_read_line · ocr_read_trials
ocr)ocr_font · ocr_render_text · ocr_segment_line · ocr_features · ocr_train · ocr_read_line · ocr_score · ocr_read_trials
*Provenance: ocrclassic.py — SYMBOL operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.