Event Transformer+. A multi-purpose solution for efficient event data processing
Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current topperforming methods often ignore specific event-data...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2023-09 |
---|---|
Hauptverfasser: | , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Sabater, Alberto Montesano, Luis Murillo, Ana C |
description | Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current topperforming methods often ignore specific event-data properties, leading to the development of generic but computationally expensive algorithms, while event-aware methods do not perform as well. We propose Event Transformer+, that improves our seminal work EvT with a refined patch-based event representation and a more robust backbone to achieve more accurate results, while still benefiting from event-data sparsity to increase its efficiency. Additionally, we show how our system can work with different data modalities and propose specific output heads, for event-stream classification (i.e. action recognition) and per-pixel predictions (dense depth estimation). Evaluation results show better performance to the state-of-the-art while requiring minimal computation resources, both on GPU and CPU. |
doi_str_mv | 10.48550/arxiv.2211.12222 |
format | Article |
fullrecord | <record><control><sourceid>proquest_arxiv</sourceid><recordid>TN_cdi_arxiv_primary_2211_12222</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2739284800</sourcerecordid><originalsourceid>FETCH-LOGICAL-a952-ae8c7d39e8ab2fa13353e7e63226285c43337595db6588b09aebaf4a029a8f683</originalsourceid><addsrcrecordid>eNotj0FLw0AQhRdBsNT-AE8ueJTEzUw22RxLqVYo6KH3MElnZUubxN2k6L83TZ3LXL73eJ8QD4mKU6O1eiH_484xQJLECYx3I2aAmEQmBbgTixAOSinIctAaZ-JzfeamlztPTbCtP7F_juVSnoZj76Ju8F0bWIb2OPSubeRISLbW1e4S4im6p55k59uaQ3DN1724tXQMvPj_c7F7Xe9Wm2j78fa-Wm4jKjRExKbO91iwoQosJYgaOecMATIwuk4RMdeF3leZNqZSBXFFNiUFBRmbGZyLx2vtpFt23p3I_5YX7XLSHomnKzFu-x449OWhHXwzbiohxwJMapTCP3G9Wpw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2739284800</pqid></control><display><type>article</type><title>Event Transformer+. A multi-purpose solution for efficient event data processing</title><source>arXiv.org</source><source>Free E- Journals</source><creator>Sabater, Alberto ; Montesano, Luis ; Murillo, Ana C</creator><creatorcontrib>Sabater, Alberto ; Montesano, Luis ; Murillo, Ana C</creatorcontrib><description>Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current topperforming methods often ignore specific event-data properties, leading to the development of generic but computationally expensive algorithms, while event-aware methods do not perform as well. We propose Event Transformer+, that improves our seminal work EvT with a refined patch-based event representation and a more robust backbone to achieve more accurate results, while still benefiting from event-data sparsity to increase its efficiency. Additionally, we show how our system can work with different data modalities and propose specific output heads, for event-stream classification (i.e. action recognition) and per-pixel predictions (dense depth estimation). Evaluation results show better performance to the state-of-the-art while requiring minimal computation resources, both on GPU and CPU.</description><identifier>EISSN: 2331-8422</identifier><identifier>DOI: 10.48550/arxiv.2211.12222</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Computer Science - Computer Vision and Pattern Recognition ; Data processing ; Temporal resolution ; Transformers</subject><ispartof>arXiv.org, 2023-09</ispartof><rights>2023. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,778,782,883,27912</link.rule.ids><backlink>$$Uhttps://doi.org/10.48550/arXiv.2211.12222$$DView paper in arXiv$$Hfree_for_read</backlink><backlink>$$Uhttps://doi.org/10.1109/TPAMI.2023.3311336$$DView published paper (Access to full text may be restricted)$$Hfree_for_read</backlink></links><search><creatorcontrib>Sabater, Alberto</creatorcontrib><creatorcontrib>Montesano, Luis</creatorcontrib><creatorcontrib>Murillo, Ana C</creatorcontrib><title>Event Transformer+. A multi-purpose solution for efficient event data processing</title><title>arXiv.org</title><description>Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current topperforming methods often ignore specific event-data properties, leading to the development of generic but computationally expensive algorithms, while event-aware methods do not perform as well. We propose Event Transformer+, that improves our seminal work EvT with a refined patch-based event representation and a more robust backbone to achieve more accurate results, while still benefiting from event-data sparsity to increase its efficiency. Additionally, we show how our system can work with different data modalities and propose specific output heads, for event-stream classification (i.e. action recognition) and per-pixel predictions (dense depth estimation). Evaluation results show better performance to the state-of-the-art while requiring minimal computation resources, both on GPU and CPU.</description><subject>Algorithms</subject><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Data processing</subject><subject>Temporal resolution</subject><subject>Transformers</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><sourceid>GOX</sourceid><recordid>eNotj0FLw0AQhRdBsNT-AE8ueJTEzUw22RxLqVYo6KH3MElnZUubxN2k6L83TZ3LXL73eJ8QD4mKU6O1eiH_484xQJLECYx3I2aAmEQmBbgTixAOSinIctAaZ-JzfeamlztPTbCtP7F_juVSnoZj76Ju8F0bWIb2OPSubeRISLbW1e4S4im6p55k59uaQ3DN1724tXQMvPj_c7F7Xe9Wm2j78fa-Wm4jKjRExKbO91iwoQosJYgaOecMATIwuk4RMdeF3leZNqZSBXFFNiUFBRmbGZyLx2vtpFt23p3I_5YX7XLSHomnKzFu-x449OWhHXwzbiohxwJMapTCP3G9Wpw</recordid><startdate>20230903</startdate><enddate>20230903</enddate><creator>Sabater, Alberto</creator><creator>Montesano, Luis</creator><creator>Murillo, Ana C</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><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230903</creationdate><title>Event Transformer+. A multi-purpose solution for efficient event data processing</title><author>Sabater, Alberto ; Montesano, Luis ; Murillo, Ana C</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a952-ae8c7d39e8ab2fa13353e7e63226285c43337595db6588b09aebaf4a029a8f683</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Algorithms</topic><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Data processing</topic><topic>Temporal resolution</topic><topic>Transformers</topic><toplevel>online_resources</toplevel><creatorcontrib>Sabater, Alberto</creatorcontrib><creatorcontrib>Montesano, Luis</creatorcontrib><creatorcontrib>Murillo, Ana C</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</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><collection>arXiv Computer Science</collection><collection>arXiv.org</collection><jtitle>arXiv.org</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Sabater, Alberto</au><au>Montesano, Luis</au><au>Murillo, Ana C</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Event Transformer+. A multi-purpose solution for efficient event data processing</atitle><jtitle>arXiv.org</jtitle><date>2023-09-03</date><risdate>2023</risdate><eissn>2331-8422</eissn><abstract>Event cameras record sparse illumination changes with high temporal resolution and high dynamic range. Thanks to their sparse recording and low consumption, they are increasingly used in applications such as AR/VR and autonomous driving. Current topperforming methods often ignore specific event-data properties, leading to the development of generic but computationally expensive algorithms, while event-aware methods do not perform as well. We propose Event Transformer+, that improves our seminal work EvT with a refined patch-based event representation and a more robust backbone to achieve more accurate results, while still benefiting from event-data sparsity to increase its efficiency. Additionally, we show how our system can work with different data modalities and propose specific output heads, for event-stream classification (i.e. action recognition) and per-pixel predictions (dense depth estimation). Evaluation results show better performance to the state-of-the-art while requiring minimal computation resources, both on GPU and CPU.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><doi>10.48550/arxiv.2211.12222</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2023-09 |
issn | 2331-8422 |
language | eng |
recordid | cdi_arxiv_primary_2211_12222 |
source | arXiv.org; Free E- Journals |
subjects | Algorithms Computer Science - Computer Vision and Pattern Recognition Data processing Temporal resolution Transformers |
title | Event Transformer+. A multi-purpose solution for efficient event data processing |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-16T00%3A52%3A01IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_arxiv&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Event%20Transformer+.%20A%20multi-purpose%20solution%20for%20efficient%20event%20data%20processing&rft.jtitle=arXiv.org&rft.au=Sabater,%20Alberto&rft.date=2023-09-03&rft.eissn=2331-8422&rft_id=info:doi/10.48550/arxiv.2211.12222&rft_dat=%3Cproquest_arxiv%3E2739284800%3C/proquest_arxiv%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2739284800&rft_id=info:pmid/&rfr_iscdi=true |