Multimodal Data Dynamic Compression Algorithm Based on Semantic Importance

With the exponential growth of multimodal data in IoT, efficiently compressing data while preserving semantic integrity has become a critical challenge. However, due to the limitations of existing single-modality and non-dynamic approaches, semantic compression still suffers from low compression eff...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE International Conference on Communications workshops S. 1694 - 1698
Hauptverfasser: Wei, Shuangying, Feng, Chunyan, Guo, Caili, Zhang, Biling
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 08.06.2025
Schlagworte:
ISSN:2694-2941
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract With the exponential growth of multimodal data in IoT, efficiently compressing data while preserving semantic integrity has become a critical challenge. However, due to the limitations of existing single-modality and non-dynamic approaches, semantic compression still suffers from low compression efficiency and incomplete semantic preservation. To address these limitations, this paper proposes a multimodal data compression algorithm based on dynamic semantic importance. Specifically, we improve the traditional text importance measurement algorithm TF-IDF by introducing a feature gradient-driven dynamic semantic importance measurement method. Combined with task-oriented weight allocation, it precisely quantifies the contribution of semantic features to task objectives. Furthermore, an compression threshold optimization algorithm is designed to dynamically balance semantic fidelity and resource constraints during compression. Simulation results demonstrate that the proposed compression method achieves a mean squared error (MSE) of 0.12 while compressing data volume by 50%. Under AWGN channels (SNR > 10 dB), low-threshold models (w 0 < 0.4) maintain stable MSE below 0.12 with downstream flood detection accuracy reaching 95.8%.
AbstractList With the exponential growth of multimodal data in IoT, efficiently compressing data while preserving semantic integrity has become a critical challenge. However, due to the limitations of existing single-modality and non-dynamic approaches, semantic compression still suffers from low compression efficiency and incomplete semantic preservation. To address these limitations, this paper proposes a multimodal data compression algorithm based on dynamic semantic importance. Specifically, we improve the traditional text importance measurement algorithm TF-IDF by introducing a feature gradient-driven dynamic semantic importance measurement method. Combined with task-oriented weight allocation, it precisely quantifies the contribution of semantic features to task objectives. Furthermore, an compression threshold optimization algorithm is designed to dynamically balance semantic fidelity and resource constraints during compression. Simulation results demonstrate that the proposed compression method achieves a mean squared error (MSE) of 0.12 while compressing data volume by 50%. Under AWGN channels (SNR > 10 dB), low-threshold models (w 0 < 0.4) maintain stable MSE below 0.12 with downstream flood detection accuracy reaching 95.8%.
Author Guo, Caili
Zhang, Biling
Wei, Shuangying
Feng, Chunyan
Author_xml – sequence: 1
  givenname: Shuangying
  surname: Wei
  fullname: Wei, Shuangying
  email: shuangying@bupt.edu.cn
  organization: Beijing University of Posts and Telecommunications,Beijing,China,100876
– sequence: 2
  givenname: Chunyan
  surname: Feng
  fullname: Feng, Chunyan
  email: cyfeng@bupt.edu.cn
  organization: Beijing University of Posts and Telecommunications,Beijing,China,100876
– sequence: 3
  givenname: Caili
  surname: Guo
  fullname: Guo, Caili
  email: guocaili@bupt.edu.cn
  organization: Beijing University of Posts and Telecommunications,Beijing,China,100876
– sequence: 4
  givenname: Biling
  surname: Zhang
  fullname: Zhang, Biling
  email: bilingzhang@bupt.edu.cn
  organization: Beijing University of Posts and Telecommunications,Beijing,China,100876
BookMark eNo1jz1PwzAURQ0CiVL6DxiyMaXYz7ETjyXlI6iIgUqM1XNsU0McR3EY-u-pBExXuufqSPeSnPWxt4TcMLpkjKrbpq7f4_iV9nFIspRlsQQK4siYBAB-QhaqVBXnTCgJRXVKZiBVkYMq2AVZpPRJKeWsqipZzMjzy3c3-RANdtkaJ8zWhx6Db7M6hmG0KfnYZ6vuI45-2ofsDpM12bF6swH76bhrwhDHCfvWXpFzh12yi7-ck-3D_bZ-yjevj0292uRe8Sl3unJKFAKRUaaVMUpQilzrEpyG1ggmHKfOgKGl4K7UVGp00HKjwElh-Zxc_2q9tXY3jD7geNj9v-c_ertU1w
ContentType Conference Proceeding
DBID 6IE
6IL
CBEJK
RIE
RIL
DOI 10.1109/ICCWorkshops67674.2025.11162223
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Xplore POP ALL
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP All) 1998-Present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Engineering
EISBN 9798331596248
EISSN 2694-2941
EndPage 1698
ExternalDocumentID 11162223
Genre orig-research
GrantInformation_xml – fundername: National Natural Science Foundation of China
  funderid: 10.13039/501100001809
GroupedDBID 6IE
6IF
6IH
6IK
6IL
6IN
AAJGR
AAWTH
ABLEC
ADZIZ
ALMA_UNASSIGNED_HOLDINGS
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CBEJK
CHZPO
IEGSK
IPLJI
OCL
RIE
RIL
ID FETCH-LOGICAL-i93t-fb8f9545aa101b9dd9500a3bb72fb2cd515f30fd2d0753f7b06baf2c3d92f65e3
IEDL.DBID RIE
IngestDate Wed Oct 01 07:05:14 EDT 2025
IsPeerReviewed false
IsScholarly false
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i93t-fb8f9545aa101b9dd9500a3bb72fb2cd515f30fd2d0753f7b06baf2c3d92f65e3
PageCount 5
ParticipantIDs ieee_primary_11162223
PublicationCentury 2000
PublicationDate 2025-June-8
PublicationDateYYYYMMDD 2025-06-08
PublicationDate_xml – month: 06
  year: 2025
  text: 2025-June-8
  day: 08
PublicationDecade 2020
PublicationTitle IEEE International Conference on Communications workshops
PublicationTitleAbbrev ICC Workshops
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0003188864
Score 1.911114
Snippet With the exponential growth of multimodal data in IoT, efficiently compressing data while preserving semantic integrity has become a critical challenge....
SourceID ieee
SourceType Publisher
StartPage 1694
SubjectTerms Accuracy
Compression algorithms
Conferences
Heuristic algorithms
multimodal compression
Optimization
Resource management
Semantic communication
semantic measurement
Semantics
Signal to noise ratio
Simulation
Title Multimodal Data Dynamic Compression Algorithm Based on Semantic Importance
URI https://ieeexplore.ieee.org/document/11162223
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LTwIxEG6EGKMXXxh8pgcTTwtrd7ftHhUk4oGQyIEb6WMqJOwugcXfb9sF1IMHb03TtM10Jt9kOt8MQvdACQBRPAhBOEpOTAPJEhEQo5TVMIeZ3DebYIMBH4_T4Yas7rkwAOCTz6Dlhv4vXxdq7UJlbWuX1OFZDdUYoxVZaxdQscrJOY0P0MOmjma73-m4iPNqWixWri6ZC6GQpLXd5Vc_FQ8nveN_XuQENb6JeXi4g5xTtAf5GTr6UVPwHL15Sm1WaDHHXVEK3K16zmNn-VXSa46f5h_FclZOM_xsUUxjO_UOmRWyXdfPvEtuj2qgUe9l1HkNNg0TglkalYGR3KTWIxLC2plMtU6TMBSRlIwYSZS2rouJQqOJtn5CZJgMqRSGqEinxNAEogtUz4scmggrBpQLrWisTMyESpWKgUvz6AhWYPQlajipTBZVSYzJViBXf8xfo0Mne59jxW9QvVyu4Rbtq89ytlre-Yf8AtLwoKU
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LTwIxEJ4oGh8XXxjf9mDiaWHt7na7RwUJKBISOXAjfQoJ7BJY_P22XUA9ePDWNE3bzHTyTabzzQDcKYKVwoJ6vmKWkhMSj8cR87AWwrwwi5nUNZuIOx3a7yfdJVndcWGUUi75TFXs0P3ly0wsbKisauySWDzbhK0oDLFf0LXWIRXzPCkl4Q7cLytpVlu1mo05z4fZdG4rk9kgCo4qq31-dVRxgNI4-OdVDqH8Tc1D3TXoHMGGSo9h_0dVwRN4caTaSSbZGNVZzlC96DqPrO0Xaa8pehx_ZLNRPpygJ4NjEpmpdzUxYjbrWhPnlJujytBrPPdqTW_ZMsEbJUHuaU51Ynwixoyl8UTKJPJ9FnAeY82xkMZ50YGvJZbGUwh0zH3CmcYikAnWJFLBKZTSLFVngESsCGVSkFDoMGYiESJUlOsHS7FSWp5D2UplMC2KYgxWArn4Y_4Wdpu9t_ag3eq8XsKe1YPLuKJXUMpnC3UN2-IzH81nN06pX1rWo-w
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=IEEE+International+Conference+on+Communications+workshops&rft.atitle=Multimodal+Data+Dynamic+Compression+Algorithm+Based+on+Semantic+Importance&rft.au=Wei%2C+Shuangying&rft.au=Feng%2C+Chunyan&rft.au=Guo%2C+Caili&rft.au=Zhang%2C+Biling&rft.date=2025-06-08&rft.pub=IEEE&rft.eissn=2694-2941&rft.spage=1694&rft.epage=1698&rft_id=info:doi/10.1109%2FICCWorkshops67674.2025.11162223&rft.externalDocID=11162223