{"id":"https://openalex.org/W4392908934","doi":"https://doi.org/10.1109/icassp48485.2024.10446532","title":"Conformer is All You Need for Visual Speech Recognition","display_name":"Conformer is All You Need for Visual Speech Recognition","publication_year":2024,"publication_date":"2024-03-18","ids":{"openalex":"https://openalex.org/W4392908934","doi":"https://doi.org/10.1109/icassp48485.2024.10446532"},"language":"en","primary_location":{"id":"doi:10.1109/icassp48485.2024.10446532","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10446532","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5073775071","display_name":"Oscar Chang","orcid":"https://orcid.org/0000-0002-4336-7545"},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Oscar Chang","raw_affiliation_strings":["Google, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google, Inc","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5110698977","display_name":"Hank Liao","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hank Liao","raw_affiliation_strings":["Google, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google, Inc","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023556869","display_name":"Dmitriy Serdyuk","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Dmitriy Serdyuk","raw_affiliation_strings":["Google, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google, Inc","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108492825","display_name":"Ankit Shah","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ankit Shah","raw_affiliation_strings":["Google, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google, Inc","institution_ids":["https://openalex.org/I1291425158"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5005881531","display_name":"Olivier Siohan","orcid":null},"institutions":[{"id":"https://openalex.org/I1291425158","display_name":"Google (United States)","ror":"https://ror.org/00njsd438","country_code":"US","type":"company","lineage":["https://openalex.org/I1291425158","https://openalex.org/I4210128969"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Olivier Siohan","raw_affiliation_strings":["Google, Inc"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Google, Inc","institution_ids":["https://openalex.org/I1291425158"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I1291425158"],"apc_list":null,"apc_paid":null,"fwci":9.5944,"has_fulltext":false,"cited_by_count":23,"citation_normalized_percentile":{"value":0.98599468,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"10136","last_page":"10140"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.48559999465942383,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10860","display_name":"Speech and Audio Processing","score":0.48559999465942383,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.657133936882019},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.5015718936920166},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.4836665093898773},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.40337449312210083}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.657133936882019},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.5015718936920166},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4836665093898773},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40337449312210083}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp48485.2024.10446532","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp48485.2024.10446532","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":27,"referenced_works":["https://openalex.org/W1614298861","https://openalex.org/W1686810756","https://openalex.org/W2890952074","https://openalex.org/W2891205112","https://openalex.org/W2972756321","https://openalex.org/W3006974783","https://openalex.org/W3016176723","https://openalex.org/W3097777922","https://openalex.org/W3160681319","https://openalex.org/W3162293946","https://openalex.org/W3196861523","https://openalex.org/W4221153068","https://openalex.org/W4221161761","https://openalex.org/W4224923231","https://openalex.org/W4287116734","https://openalex.org/W4288327876","https://openalex.org/W4385245566","https://openalex.org/W4390040710","https://openalex.org/W6636510571","https://openalex.org/W6637373629","https://openalex.org/W6735927292","https://openalex.org/W6754420807","https://openalex.org/W6764045775","https://openalex.org/W6798160016","https://openalex.org/W6810168380","https://openalex.org/W6810673746","https://openalex.org/W6859663106"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052","https://openalex.org/W2382290278","https://openalex.org/W3204019825"],"abstract_inverted_index":{"Visual":[0],"speech":[1,77,134],"recognition":[2,135],"models":[3,147],"extract":[4,72],"visual":[5,18,66,85,97,104,133],"features":[6,75],"in":[7,113],"a":[8,17,21,54,95,102,108,126],"hierarchical":[9],"manner.":[10],"At":[11,36],"the":[12,28,32,37,47,51,65,69,137,143],"lower":[13,114],"level,":[14,39],"there":[15,40],"is":[16,41],"front-end":[19,52,67,105],"with":[20,107],"limited":[22],"temporal":[23,56],"receptive":[24,57],"field":[25],"that":[26,44,83,101],"processes":[27],"raw":[29],"pixels":[30],"depicting":[31],"lips":[33],"or":[34],"faces.":[35],"higher":[38],"an":[42],"encoder":[43,111],"attends":[45],"to":[46,71,94],"embeddings":[48],"produced":[49],"by":[50],"over":[53],"large":[55],"field.":[58],"Previous":[59],"work":[60,81],"has":[61],"focused":[62],"on":[63,136],"improving":[64],"of":[68,91,129,145],"model":[70],"more":[73,116],"useful":[74],"for":[76,132],"recognition.":[78],"Surprisingly,":[79],"our":[80],"shows":[82],"complex":[84],"front-ends":[86],"are":[87],"not":[88],"necessary.":[89],"Instead":[90],"allocating":[92],"resources":[93],"sophisticated":[96],"front-end,":[98],"we":[99],"find":[100],"linear":[103],"paired":[106],"larger":[109],"Conformer":[110],"results":[112],"latency,":[115],"efficient":[117],"memory":[118],"usage,":[119],"and":[120],"improved":[121],"WER":[122,131],"performance.":[123],"We":[124],"achieve":[125],"new":[127],"state-of-the-art":[128],"12.8%":[130],"TED":[138],"LRS3":[139],"dataset,":[140],"which":[141],"rivals":[142],"performance":[144],"audio-only":[146],"from":[148],"just":[149],"four":[150],"years":[151],"ago.":[152]},"counts_by_year":[{"year":2026,"cited_by_count":2},{"year":2025,"cited_by_count":13},{"year":2024,"cited_by_count":8}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}