[{"@type":"PropertyValue","name":"Data size","value":"101 people, 4,538 images"},{"@type":"PropertyValue","name":"Collecting environment","value":"A4 paper"},{"@type":"PropertyValue","name":"Device","value":"scanner"},{"@type":"PropertyValue","name":"Photographic angle","value":"eye-level angle"},{"@type":"PropertyValue","name":"Data format","value":"the image data format is .jpg, the annotation file format is .json"},{"@type":"PropertyValue","name":"Data content","value":"including social livelihood, entertainment, tour, sport, movie, composition and other fields"},{"@type":"PropertyValue","name":"Annotation content","value":"character-level rectangular bounding box annotation and text transcription; line-level rectangular bounding box annotation and text transcription"},{"@type":"PropertyValue","name":"Accuracy","value":"the error bound of each vertex of rectangular bounding box is within 3 pixels, which is a qualified annotation, the accuracy of bounding boxes is not less than 97%; the characters transcription accuracy is not less than 97%"}]
{"id":1087,"datatype":"1","titleimg":"https://res.datatang.com/asset/productNew/APY201029001.png?Expires=2007353688&OSSAccessKeyId=LTAI5tQwXnJZbubgVfVa1ep9&Signature=Ctg%2BRaUjv/UviDD6xKiFSheIWqY%3D","type1":"147","type1str":null,"type2":"150","type2str":null,"dataname":"Japanese Handwriting OCR Dataset – 4,538 Handwritten Text Images","datazy":[{"title":"Data size","desc":"Data size","content":"101 people, 4,538 images"},{"title":"Collecting environment","desc":"Collecting environment","content":"A4 paper"},{"title":"Device","desc":"Device","content":"scanner"},{"title":"Photographic angle","desc":"Photographic angle","content":"eye-level angle"},{"title":"Data format","desc":"Data format","content":"the image data format is .jpg, the annotation file format is .json"},{"title":"Data content","desc":"Data content","content":"including social livelihood, entertainment, tour, sport, movie, composition and other fields"},{"title":"Annotation content","desc":"Annotation content","content":"character-level rectangular bounding box annotation and text transcription; line-level rectangular bounding box annotation and text transcription"},{"title":"Accuracy","desc":"Accuracy","content":"the error bound of each vertex of rectangular bounding box is within 3 pixels, which is a qualified annotation, the accuracy of bounding boxes is not less than 97%; the characters transcription accuracy is not less than 97%"}],"datatag":"Japanese,Handwriting,OCR","technologydoc":null,"downurl":null,"datainfo":null,"standard":null,"dataylurl":null,"flag":null,"publishtime":null,"createby":null,"createtime":null,"ext1":null,"samplestoreloc":null,"hosturl":null,"datasize":null,"industryPlan":null,"keyInformation":null,"samplePresentation":[{"name":"/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/2.jpg","url":"https://bj-oss-datatang-03.oss-cn-beijing.aliyuncs.com/filesInfoUpload/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/2.jpg?Expires=4102329599&OSSAccessKeyId=LTAI8NWs2pDolLNH&Signature=Wr%2FmvyKp%2B%2Fw6OVZ7Etgqu1vTjJ0%3D","intro":"","size":0,"progress":100,"type":"jpg"},{"name":"/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/3.jpg","url":"https://bj-oss-datatang-03.oss-cn-beijing.aliyuncs.com/filesInfoUpload/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/3.jpg?Expires=4102329599&OSSAccessKeyId=LTAI8NWs2pDolLNH&Signature=Vl165NGIjihXPtcQJ15IkguuVX8%3D","intro":"","size":0,"progress":100,"type":"jpg"},{"name":"/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/1.jpg","url":"https://bj-oss-datatang-03.oss-cn-beijing.aliyuncs.com/filesInfoUpload/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/1.jpg?Expires=4102329599&OSSAccessKeyId=LTAI8NWs2pDolLNH&Signature=%2FiYAJa11zvmV3Qz0AbFvn3aVlPU%3D","intro":"","size":0,"progress":100,"type":"jpg"}],"officialSummary":"This dataset contains 4,538 Japanese handwritten text images collected from 101 individual writers, written on A4 paper. The dataset content including social livelihood, entertainment, tour, sport, movie, composition and other fields. For annotation, character-level rectangular bounding box annotation and text transcription and line-level rectangular bounding box annotation and text transcription were adopted. The dataset can be used for for training and evaluating Japanese handwriting OCR models, handwritten text recognition systems, and document understanding pipelines.","dataexampl":null,"datakeyword":["handwriting OCR dataset","handwritten OCR dataset","handwriting recognition dataset","Japanese handwriting OCR dataset"],"isDelete":null,"ids":null,"idsList":null,"datasetCode":null,"productStatus":null,"tagTypeEn":"Data Type,Language","tagTypeZh":null,"website":null,"samplePresentationList":null,"datazyList":null,"keyInformationList":null,"dataexamplList":null,"bgimg":null,"datazyScriptList":null,"datakeywordListString":null,"sourceShowPage":"ocr","dataShowType":"[{\"code\":\"0\",\"language\":\"ZH\"},{\"code\":\"1\",\"language\":\"ZH\"},{\"code\":\"2\",\"language\":\"EN,JP,PT,DE,KO,FR,ES\"},{\"code\":\"3\",\"language\":\"EN\"}]","productNameEn":"101 People - 4,538 Images Japanese Handwriting OCR Data","BGimg":"","voiceBg":["/shujutang/static/image/comm/audio_bg.webp","/shujutang/static/image/comm/audio_bg2.webp","/shujutang/static/image/comm/audio_bg3.webp","/shujutang/static/image/comm/audio_bg4.webp","/shujutang/static/image/comm/audio_bg5.webp"],"firstList":[{"name":"/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/5.jpg","url":"https://bj-oss-datatang-03.oss-cn-beijing.aliyuncs.com/filesInfoUpload/data/apps/damp/temp/ziptemp/APY201029001_demo1693389600574/APY201029001_demo/5.jpg?Expires=4102329599&OSSAccessKeyId=LTAI8NWs2pDolLNH&Signature=jCFGBgP1FC9%2F4JimiPwL4XVNN4g%3D","intro":"","size":0,"progress":100,"type":"jpg"}]}
Japanese Handwriting OCR Dataset – 4,538 Handwritten Text Images
handwriting OCR dataset
handwritten OCR dataset
handwriting recognition dataset
Japanese handwriting OCR dataset
This dataset contains 4,538 Japanese handwritten text images collected from 101 individual writers, written on A4 paper. The dataset content including social livelihood, entertainment, tour, sport, movie, composition and other fields. For annotation, character-level rectangular bounding box annotation and text transcription and line-level rectangular bounding box annotation and text transcription were adopted. The dataset can be used for for training and evaluating Japanese handwriting OCR models, handwritten text recognition systems, and document understanding pipelines.
This is a paid datasets for commercial use, research purpose and more. Licensed ready made datasets help jump-start AI projects.
Specifications
Data size
101 people, 4,538 images
Collecting environment
A4 paper
Device
scanner
Photographic angle
eye-level angle
Data format
the image data format is .jpg, the annotation file format is .json
Data content
including social livelihood, entertainment, tour, sport, movie, composition and other fields
Annotation content
character-level rectangular bounding box annotation and text transcription; line-level rectangular bounding box annotation and text transcription
Accuracy
the error bound of each vertex of rectangular bounding box is within 3 pixels, which is a qualified annotation, the accuracy of bounding boxes is not less than 97%; the characters transcription accuracy is not less than 97%
Nexdata provides OCR datasets covering a wide range of data types, including documents, handwriting, invoices, test papers, and forms. The datasets cover diverse languages, layouts, text styles, and real-world scenarios to support OCR model training, text recognition, document analysis, and other document AI applications.
Can Nexdata customize OCR datasets for specific requirements?
Yes. If our off-the-shelf OCR datasets do not fully meet your requirements, Nexdata provides flexible custom data collection, annotation, and quality control services. We can customize datasets based on target languages, document types, handwriting styles, layouts, image conditions, data volume, and annotation specifications for specific OCR applications.
How does Nexdata ensure the quality and scalability of OCR datasets?
Nexdata applies multi-stage quality control throughout data collection, annotation, validation, and delivery. With extensive resources covering documents, handwriting, invoices, test papers, and forms, we can support both large-scale OCR projects and specialized datasets with customized formats, annotation requirements, and quality standards.
Specifications
Sample
Recommended Dataset
