InoteveraGaidhi rinotevera
Kutsanangurwa kweWord Error Rate
AI Yenyaya
Nhungamiro yehunyanzvi
Chiyero chekutadza kwemavara (CER) uye chiyero chekutadza kwemashoko (WER) chinoenzanisa zvinyorwa zvakazivikanwa ne transcript yekureferensi uchishandisa zvikanganiso zveedit-distance.
Teerera kugwaro iriAI-yakarondedzerwa · ~3 min
CER inoverenga kutsiva kwemavara, kubvisa, uye kuwedzera maererano nemavara ekureferensi; WER inoshanda pfungwa imwechete kumashoko. Mamaki anobva pakugadzirisa transcription, tokenization, uye marongerwo egwaro, saka metric inofanira kutaurwa nemirairo yayo yekuyera uye haifaniri kutsiva kuongorora kwezvikanganiso zvakakosha.
OCR output can be scored by aligning it with a reference transcript and counting edit operations. Character error rate divides character substitutions, deletions, and insertions by the number of characters in the reference. Word error rate uses the same edit-distance components at the word level and divides by the number of reference words. These measures provide a reproducible way to compare systems on the same labeled examples, but their meaning depends on how text is normalized and tokenized. For example, a pipeline might lowercase both strings, remove repeated spaces, or ignore punctuation before computing CER. A WER evaluation must define what counts as a word, especially across languages, scripts, contractions, and writing systems without spaces. Different normalization rules can produce different scores even when the OCR output is unchanged. The denominator is the reference length, and insertions can make either rate exceed 1.0. A low overall score can also hide a critical error in one account number or a missed line in a table. Report the reference source, preprocessing, tokenization, scoring implementation, and whether you aggregate per document or globally. Use CER for character-level fidelity and WER for word-level text, then inspect errors for the task. For document extraction, also measure field-level correctness and layout or reading-order quality. Do not compare published OCR scores unless their datasets and normalization rules are comparable. No single metric captures every kind of document error or downstream risk.
Chikwata chinofananidza mavhezheni maviri eOCR pane seti yakatarwa yeground-truth uye chinotaura CER mushure mekushandisa Unicode ne whitespace normalization imwechete.
Muongorori wegwaro anoshandisa WER kuongorora zvinyorwa zvepasipeji asi zvakasiyana anoongorora kana mazuva, huwandu, uye zviziviso zvakaverengwa nemazvo.
Mutsvakurudzi anotaura nhamba dzekuisa, kubvisa, uye kutsiva pamwe neCER kuti vaverengi vawone kuti chiyero chachinja sei.
A multilingual OCR evaluation inoratidza maitiro ayo pakugadzirisa punctuation, whitespace, uye script-specific tokenization isati yaitwa kuenzanisa masisitimu.
Maitiro eOCR acharamba achishandisa CER neWER nekuti ari nyore kudzokororwa, nepo zviyero zvinoenderana nebasa zvichidiwa kumatafura, mafomu, uye minda ine pesvedzero huru. Mamodheru ari nani ezvinyorwa haabvise kudiwa kwekugadzirisa kuenzanisa kana kuongorora mhando dzezvikanganiso. Buritsai script yekuongorora pamwe nepolicy yekureferensi kuti kuenzanisa kurambe kuchinzwisisika sezvo masisitimu eOCR achichinja. Kana mapatani ezvinyorwa achichinja, ongororai kusarudzwa kwemienzaniso uye zviyero zvemhando yenzvimbo usati waenzanisa mhedzisiro itsva nemapoinzi ekare. Ratidzirai kusava nechokwadi kana mienzaniso yekusimbisa iri diki.
Tsamba yemahara
Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.
Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.
Zviedze iwe pachako
Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.
Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo
Character error rate (CER) uye word error rate (WER) zvinofananidza zvinyorwa zvakazivikanwa ne transcript yekureferensi vachishandisa zvikanganiso zveedit-distance. CER inoverenga kutsiva, kubviswa, uye kuwedzerwa kwemavara zvichienzaniswa ne mavara ekureferensi; WER inoshanda pfungwa imwechete kumashoko. Zvigadzirwa zvinoenderana nekugadzirisa transcription, tokenization, uye marongerwo ezvinyorwa, saka metric inofanira kuratidzwa pamwe nemirau yekuyedzwa uye haifaniri kutsiva kuongorora kwezvikanganiso zvakakosha.
CER inoshandisa mashandiro eedit pamusoro pekuverengwa kwemavara ekureferensi.
WER inoshandisa kuwiriranisa pamazwi uye denominator yemazwi ekureferensi.
Kuisa kunogona kuita kuti S+D+I zvikure kupfuura huwandu hwezvikamu zvekureferensi.
Unicode normalization, punctuation handling, uye whitespace processing zvinogona kuchinja kuwirirana kwemavara uye CER inobuda.
Mavhoti akabatanidzwa anogona kuvanzwa zvikanganiso zvinodhura zvakanyanya muminda chaiyo.
Ramba uchidzidza
Mamwe madhairekitori akasarudzirwa nyaya iyi
InoteveraGaidhi rinotevera
Kutsanangurwa kweWord Error Rate
AI Yenyaya