Graduate School and Research Center in Digital Sciences

The open-set problem in acoustic scene classification

Battaglino, Daniele; Lepauloux, Ludovick; Evans, Nicholas

IWAENC 2016, 15th International Workshop on Acoustic Signal Enhancement, September 13-16, 2016, Xi'an, China

Acoustic scene classification (ASC) has attracted growing research interest in recent years. Whereas the previous work has investigated closed-set classification scenarios, the predominant ASC application is open-set in nature. The contributions of the paper are (i) the first investigation of ASC in an open-set scenario, (ii) the formulation of open-set ASC as a detection problem, (iii) a classifier tailored to the open-set scenario and (iv) a new assessment protocol and metric. Experiments show that, despite the challenge of open-set ASC, reliable performance is achieved with the support vector data description classifier for varying levels of openness.

Document Doi Bibtex

Title:The open-set problem in acoustic scene classification
Keywords:Acoustic scene classification, open-set, support vector data description
Department:Digital Security
Eurecom ref:4954
Copyright: © 2016 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
Bibtex: @inproceedings{EURECOM+4954, doi = {}, year = {2016}, title = {{T}he open-set problem in acoustic scene classification}, author = {{B}attaglino, {D}aniele and {L}epauloux, {L}udovick and {E}vans, {N}icholas}, booktitle = {{IWAENC} 2016, 15th {I}nternational {W}orkshop on {A}coustic {S}ignal {E}nhancement, {S}eptember 13-16, 2016, {X}i'an, {C}hina\&\#13;\&\#10;}, address = {{X}i?an, {CHINA}}, month = {09}, url = {} }
See also: