Automated Focal Loss for Image based Object Detection

Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manual...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2019-04
Main Authors: Weber, Michael, Fürst, Michael, Zöllner, J Marius
Format: Article
Language:English
Subjects:
Algorithms
Automation
Image detection
Object recognition
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by
cites
container_end_page
container_issue
container_start_page
container_title arXiv.org
container_volume
creator Weber, Michael
Fürst, Michael
Zöllner, J Marius
description Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manually tuning this hyperparameter for each training task is highly time-consuming. With automated focal loss we introduce a new loss function which substitutes this hyperparameter by a parameter that is automatically adapted during the training progress and controls the amount of focusing on hard training examples. We show on the COCO benchmark that this leads to an up to 30% faster training convergence. We further introduced a focal regression loss which on the more challenging task of 3D vehicle detection outperforms other loss functions by up to 1.8 AOS and can be used as a value range independent metric for regression.
format article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2212339389</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2212339389</sourcerecordid><originalsourceid>FETCH-proquest_journals_22123393893</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQwdSwtyc9NLElNUXDLT07MUfDJLy5WSMsvUvDMTUxPVUhKLAZK-SdlpSaXKLiklgCpzPw8HgbWtMSc4lReKM3NoOzmGuLsoVtQlF9YmlpcEp-VX1qUB5SKNzIyBFpvaWxhaUycKgA1qTQ3</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2212339389</pqid></control><display><type>article</type><title>Automated Focal Loss for Image based Object Detection</title><source>Publicly Available Content (ProQuest)</source><creator>Weber, Michael ; Fürst, Michael ; Zöllner, J Marius</creator><creatorcontrib>Weber, Michael ; Fürst, Michael ; Zöllner, J Marius</creatorcontrib><description>Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manually tuning this hyperparameter for each training task is highly time-consuming. With automated focal loss we introduce a new loss function which substitutes this hyperparameter by a parameter that is automatically adapted during the training progress and controls the amount of focusing on hard training examples. We show on the COCO benchmark that this leads to an up to 30% faster training convergence. We further introduced a focal regression loss which on the more challenging task of 3D vehicle detection outperforms other loss functions by up to 1.8 AOS and can be used as a value range independent metric for regression.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Automation ; Image detection ; Object recognition ; Training</subject><ispartof>arXiv.org, 2019-04</ispartof><rights>2019. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><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://www.proquest.com/docview/2212339389?pq-origsite=primo$$EHTML$$P50$$Gproquest$$Hfree_for_read</linktohtml><link.rule.ids>776,780,25731,36989,44566</link.rule.ids></links><search><creatorcontrib>Weber, Michael</creatorcontrib><creatorcontrib>Fürst, Michael</creatorcontrib><creatorcontrib>Zöllner, J Marius</creatorcontrib><title>Automated Focal Loss for Image based Object Detection</title><title>arXiv.org</title><description>Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manually tuning this hyperparameter for each training task is highly time-consuming. With automated focal loss we introduce a new loss function which substitutes this hyperparameter by a parameter that is automatically adapted during the training progress and controls the amount of focusing on hard training examples. We show on the COCO benchmark that this leads to an up to 30% faster training convergence. We further introduced a focal regression loss which on the more challenging task of 3D vehicle detection outperforms other loss functions by up to 1.8 AOS and can be used as a value range independent metric for regression.</description><subject>Algorithms</subject><subject>Automation</subject><subject>Image detection</subject><subject>Object recognition</subject><subject>Training</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><sourceid>PIMPY</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mQwdSwtyc9NLElNUXDLT07MUfDJLy5WSMsvUvDMTUxPVUhKLAZK-SdlpSaXKLiklgCpzPw8HgbWtMSc4lReKM3NoOzmGuLsoVtQlF9YmlpcEp-VX1qUB5SKNzIyBFpvaWxhaUycKgA1qTQ3</recordid><startdate>20190419</startdate><enddate>20190419</enddate><creator>Weber, Michael</creator><creator>Fürst, Michael</creator><creator>Zöllner, J Marius</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20190419</creationdate><title>Automated Focal Loss for Image based Object Detection</title><author>Weber, Michael ; Fürst, Michael ; Zöllner, J Marius</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_22123393893</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Algorithms</topic><topic>Automation</topic><topic>Image detection</topic><topic>Object recognition</topic><topic>Training</topic><toplevel>online_resources</toplevel><creatorcontrib>Weber, Michael</creatorcontrib><creatorcontrib>Fürst, Michael</creatorcontrib><creatorcontrib>Zöllner, J Marius</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; Engineering Collection</collection><collection>ProQuest Central (Alumni)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>AUTh Library subscriptions: ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>ProQuest Engineering Database</collection><collection>Publicly Available Content (ProQuest)</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Weber, Michael</au><au>Fürst, Michael</au><au>Zöllner, J Marius</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Automated Focal Loss for Image based Object Detection</atitle><jtitle>arXiv.org</jtitle><date>2019-04-19</date><risdate>2019</risdate><eissn>2331-8422</eissn><abstract>Current state-of-the-art object detection algorithms still suffer the problem of imbalanced distribution of training data over object classes and background. Recent work introduced a new loss function called focal loss to mitigate this problem, but at the cost of an additional hyperparameter. Manually tuning this hyperparameter for each training task is highly time-consuming. With automated focal loss we introduce a new loss function which substitutes this hyperparameter by a parameter that is automatically adapted during the training progress and controls the amount of focusing on hard training examples. We show on the COCO benchmark that this leads to an up to 30% faster training convergence. We further introduced a focal regression loss which on the more challenging task of 3D vehicle detection outperforms other loss functions by up to 1.8 AOS and can be used as a value range independent metric for regression.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record>
fulltext fulltext
identifier EISSN: 2331-8422
ispartof arXiv.org, 2019-04
issn 2331-8422
language eng
recordid cdi_proquest_journals_2212339389
source Publicly Available Content (ProQuest)
subjects Algorithms
Automation
Image detection
Object recognition
Training
title Automated Focal Loss for Image based Object Detection
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-07T23%3A17%3A14IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Automated%20Focal%20Loss%20for%20Image%20based%20Object%20Detection&rft.jtitle=arXiv.org&rft.au=Weber,%20Michael&rft.date=2019-04-19&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2212339389%3C/proquest%3E%3Cgrp_id%3Ecdi_FETCH-proquest_journals_22123393893%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2212339389&rft_id=info:pmid/&rfr_iscdi=true