| Presentation | 2020-12-01 A study of anomalous sound detection using sound activity detection Yasuhiro Kanishima, Takashi Sudo, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In anomalous sound detection that determines the operation of the device or the quality of the product based on the sound signal generated from the device or product on the mass production line, for the purpose of improving the accuracy and reducing the amount of processing of anomalous sound detection, We propose a segmentation that detects the section in which the device operates from the sound signal using trigger detection and template matching detection. Using the data of the mass production line inspection sound of the rotating body product as a motif, the accuracy of abnormal sound detection by the autoencoder was evaluated for the data segmented by the conventional method and the proposed method. As a result, the AUC for the conventional method was 0.642 and the AUC for the proposed method was 0.997. We confirmed the effect of the accuracy improvement by the proposed method. In addition, intermittent processing for anomalous sound detection is possible, and the effect of reducing the processing amount can be expected. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Anomalous Sound / Anomaly Detection / Neural Network / Deep Learning / Activity Detection |
| Paper # | SIS2020-29 |
| Date of Issue | 2020-11-24 (SIS) |
| Conference Information | |
| Committee | SIS |
|---|---|
| Conference Date | 2020/12/1(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Smart Personal Systems, etc. |
| Chair | Noriaki Suetake(Yamaguchi Univ.) |
| Vice Chair | Tomoaki Kimura(Kanagawa Inst. of Tech.) / Naoto Sasaoka(Tottori Univ.) |
| Secretary | Tomoaki Kimura(Kindai Univ.) / Naoto Sasaoka(National Inst. of Tech., Ube College) |
| Assistant | Yukihiro Bandoh(NTT) / Soh Yoshida(Kansai Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Smart Info-Media Systems |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A study of anomalous sound detection using sound activity detection |
| Sub Title (in English) | |
| Keyword(1) | Anomalous Sound |
| Keyword(2) | Anomaly Detection |
| Keyword(3) | Neural Network |
| Keyword(4) | Deep Learning |
| Keyword(5) | Activity Detection |
| 1st Author's Name | Yasuhiro Kanishima |
| 1st Author's Affiliation | Toshiba Corporation(Toshiba) |
| 2nd Author's Name | Takashi Sudo |
| 2nd Author's Affiliation | Toshiba Corporation(Toshiba) |
| Date | 2020-12-01 |
| Paper # | SIS2020-29 |
| Volume (vol) | vol.120 |
| Number (no) | SIS-269 |
| Page | pp.pp.12-17(SIS), |
| #Pages | 6 |
| Date of Issue | 2020-11-24 (SIS) |