Training data cleaning method, portrait segmentation network training method and portrait segmentation method

The invention is suitable for the technical field of data processing, and provides a training data cleaning method, a portrait segmentation network training method and a portrait segmentation method, and the training data cleaning method comprises the steps: obtaining a first training data set; wher...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: MO YAOYANG
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator MO YAOYANG
description The invention is suitable for the technical field of data processing, and provides a training data cleaning method, a portrait segmentation network training method and a portrait segmentation method, and the training data cleaning method comprises the steps: obtaining a first training data set; wherein the first training data set comprises training images containing face information; performing face detection on the training image to obtain face labeling information and face orientation information corresponding to the face labeling information; the training images with the face orientation information not meeting the preset orientation condition in the first training data set are removed, and a second training data set is obtained; calculating a human face area proportion corresponding to the training image in the second training data set according to the human face labeling information; and the training images of which the face area proportions do not meet a preset proportion condition in the second trainin
format Patent
fullrecord <record><control><sourceid>epo_EVB</sourceid><recordid>TN_cdi_epo_espacenet_CN115424325A</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>CN115424325A</sourcerecordid><originalsourceid>FETCH-epo_espacenet_CN115424325A3</originalsourceid><addsrcrecordid>eNrjZMgNKUrMzMvMS1dISSxJVEjOSU0E83JTSzLyU3QUCvKLSoAqShSKU9NzU_NKEksy8_MU8lJLyvOLshVKYJohyhUS81Jw6IAo4GFgTUvMKU7lhdLcDIpuriHOHrqpBfnxqcUFicmpQKPjnf0MDU1NjEyMjUwdjYlRAwDwbUMv</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>patent</recordtype></control><display><type>patent</type><title>Training data cleaning method, portrait segmentation network training method and portrait segmentation method</title><source>esp@cenet</source><creator>MO YAOYANG</creator><creatorcontrib>MO YAOYANG</creatorcontrib><description>The invention is suitable for the technical field of data processing, and provides a training data cleaning method, a portrait segmentation network training method and a portrait segmentation method, and the training data cleaning method comprises the steps: obtaining a first training data set; wherein the first training data set comprises training images containing face information; performing face detection on the training image to obtain face labeling information and face orientation information corresponding to the face labeling information; the training images with the face orientation information not meeting the preset orientation condition in the first training data set are removed, and a second training data set is obtained; calculating a human face area proportion corresponding to the training image in the second training data set according to the human face labeling information; and the training images of which the face area proportions do not meet a preset proportion condition in the second trainin</description><language>chi ; eng</language><subject>CALCULATING ; COMPUTING ; COUNTING ; PHYSICS</subject><creationdate>2022</creationdate><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20221202&amp;DB=EPODOC&amp;CC=CN&amp;NR=115424325A$$EHTML$$P50$$Gepo$$Hfree_for_read</linktohtml><link.rule.ids>230,308,780,885,25563,76418</link.rule.ids><linktorsrc>$$Uhttps://worldwide.espacenet.com/publicationDetails/biblio?FT=D&amp;date=20221202&amp;DB=EPODOC&amp;CC=CN&amp;NR=115424325A$$EView_record_in_European_Patent_Office$$FView_record_in_$$GEuropean_Patent_Office$$Hfree_for_read</linktorsrc></links><search><creatorcontrib>MO YAOYANG</creatorcontrib><title>Training data cleaning method, portrait segmentation network training method and portrait segmentation method</title><description>The invention is suitable for the technical field of data processing, and provides a training data cleaning method, a portrait segmentation network training method and a portrait segmentation method, and the training data cleaning method comprises the steps: obtaining a first training data set; wherein the first training data set comprises training images containing face information; performing face detection on the training image to obtain face labeling information and face orientation information corresponding to the face labeling information; the training images with the face orientation information not meeting the preset orientation condition in the first training data set are removed, and a second training data set is obtained; calculating a human face area proportion corresponding to the training image in the second training data set according to the human face labeling information; and the training images of which the face area proportions do not meet a preset proportion condition in the second trainin</description><subject>CALCULATING</subject><subject>COMPUTING</subject><subject>COUNTING</subject><subject>PHYSICS</subject><fulltext>true</fulltext><rsrctype>patent</rsrctype><creationdate>2022</creationdate><recordtype>patent</recordtype><sourceid>EVB</sourceid><recordid>eNrjZMgNKUrMzMvMS1dISSxJVEjOSU0E83JTSzLyU3QUCvKLSoAqShSKU9NzU_NKEksy8_MU8lJLyvOLshVKYJohyhUS81Jw6IAo4GFgTUvMKU7lhdLcDIpuriHOHrqpBfnxqcUFicmpQKPjnf0MDU1NjEyMjUwdjYlRAwDwbUMv</recordid><startdate>20221202</startdate><enddate>20221202</enddate><creator>MO YAOYANG</creator><scope>EVB</scope></search><sort><creationdate>20221202</creationdate><title>Training data cleaning method, portrait segmentation network training method and portrait segmentation method</title><author>MO YAOYANG</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-epo_espacenet_CN115424325A3</frbrgroupid><rsrctype>patents</rsrctype><prefilter>patents</prefilter><language>chi ; eng</language><creationdate>2022</creationdate><topic>CALCULATING</topic><topic>COMPUTING</topic><topic>COUNTING</topic><topic>PHYSICS</topic><toplevel>online_resources</toplevel><creatorcontrib>MO YAOYANG</creatorcontrib><collection>esp@cenet</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>MO YAOYANG</au><format>patent</format><genre>patent</genre><ristype>GEN</ristype><title>Training data cleaning method, portrait segmentation network training method and portrait segmentation method</title><date>2022-12-02</date><risdate>2022</risdate><abstract>The invention is suitable for the technical field of data processing, and provides a training data cleaning method, a portrait segmentation network training method and a portrait segmentation method, and the training data cleaning method comprises the steps: obtaining a first training data set; wherein the first training data set comprises training images containing face information; performing face detection on the training image to obtain face labeling information and face orientation information corresponding to the face labeling information; the training images with the face orientation information not meeting the preset orientation condition in the first training data set are removed, and a second training data set is obtained; calculating a human face area proportion corresponding to the training image in the second training data set according to the human face labeling information; and the training images of which the face area proportions do not meet a preset proportion condition in the second trainin</abstract><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier
ispartof
issn
language chi ; eng
recordid cdi_epo_espacenet_CN115424325A
source esp@cenet
subjects CALCULATING
COMPUTING
COUNTING
PHYSICS
title Training data cleaning method, portrait segmentation network training method and portrait segmentation method
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T09%3A04%3A43IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-epo_EVB&rft_val_fmt=info:ofi/fmt:kev:mtx:patent&rft.genre=patent&rft.au=MO%20YAOYANG&rft.date=2022-12-02&rft_id=info:doi/&rft_dat=%3Cepo_EVB%3ECN115424325A%3C/epo_EVB%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true