A Fast Identification Method of Gunshot Types Based on Knowledge Distillation
To reduce the large size of a gunshot recognition network model and to improve the insufficient real-time detection in urban combat, this paper proposes a fast gunshot type recognition method based on knowledge distillation. First, the muzzle blast and the shock wave generated by the gunshot are pre...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-05-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/11/5526 |
| _version_ | 1797494200161271808 |
|---|---|
| author | Jian Li Jinming Guo Xiushan Sun Chuankun Li Lingpeng Meng |
| author_facet | Jian Li Jinming Guo Xiushan Sun Chuankun Li Lingpeng Meng |
| author_sort | Jian Li |
| collection | DOAJ |
| description | To reduce the large size of a gunshot recognition network model and to improve the insufficient real-time detection in urban combat, this paper proposes a fast gunshot type recognition method based on knowledge distillation. First, the muzzle blast and the shock wave generated by the gunshot are preprocessed, and the quality of the gunshot recognition dataset is improved using Log-Mel spectrum corresponding to these two signals. Second, a teacher network is constructed using 10 two-dimensional residual modules, and a student network is designed using depth wise separable convolution. Third, the lightweight student network is made to learn the gunshot features under the guidance of the pre-trained large-scale teacher network. Finally, the network’s accuracy, model size, and recognition time are tested using the AudioSet dataset and the NIJ Grant 2016-DN-BX-0183 gunshot dataset. The findings demonstrate that the proposed algorithm achieved 95.6% and 83.5% accuracy on the two datasets, the speed was 0.5 s faster, and the model size was reduced to 2.5 MB. The proposed method is of good practical value in the field of gunshot recognition. |
| first_indexed | 2024-03-10T01:30:56Z |
| format | Article |
| id | doaj.art-b8ce14a23b7144f688d56e9b9368d3b4 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-10T01:30:56Z |
| publishDate | 2022-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-b8ce14a23b7144f688d56e9b9368d3b42023-11-23T13:43:15ZengMDPI AGApplied Sciences2076-34172022-05-011211552610.3390/app12115526A Fast Identification Method of Gunshot Types Based on Knowledge DistillationJian Li0Jinming Guo1Xiushan Sun2Chuankun Li3Lingpeng Meng4National Key Laboratory of Electronic Testing Technology, North University of China, Taiyuan 030051, ChinaNational Key Laboratory of Electronic Testing Technology, North University of China, Taiyuan 030051, ChinaNational Key Laboratory of Electronic Testing Technology, North University of China, Taiyuan 030051, ChinaNational Key Laboratory of Electronic Testing Technology, North University of China, Taiyuan 030051, ChinaHunan Vanguard Group Co., Ltd., Changsha 410100, ChinaTo reduce the large size of a gunshot recognition network model and to improve the insufficient real-time detection in urban combat, this paper proposes a fast gunshot type recognition method based on knowledge distillation. First, the muzzle blast and the shock wave generated by the gunshot are preprocessed, and the quality of the gunshot recognition dataset is improved using Log-Mel spectrum corresponding to these two signals. Second, a teacher network is constructed using 10 two-dimensional residual modules, and a student network is designed using depth wise separable convolution. Third, the lightweight student network is made to learn the gunshot features under the guidance of the pre-trained large-scale teacher network. Finally, the network’s accuracy, model size, and recognition time are tested using the AudioSet dataset and the NIJ Grant 2016-DN-BX-0183 gunshot dataset. The findings demonstrate that the proposed algorithm achieved 95.6% and 83.5% accuracy on the two datasets, the speed was 0.5 s faster, and the model size was reduced to 2.5 MB. The proposed method is of good practical value in the field of gunshot recognition.https://www.mdpi.com/2076-3417/12/11/5526gunshotsLog-Mel spectrumknowledge distillation |
| spellingShingle | Jian Li Jinming Guo Xiushan Sun Chuankun Li Lingpeng Meng A Fast Identification Method of Gunshot Types Based on Knowledge Distillation Applied Sciences gunshots Log-Mel spectrum knowledge distillation |
| title | A Fast Identification Method of Gunshot Types Based on Knowledge Distillation |
| title_full | A Fast Identification Method of Gunshot Types Based on Knowledge Distillation |
| title_fullStr | A Fast Identification Method of Gunshot Types Based on Knowledge Distillation |
| title_full_unstemmed | A Fast Identification Method of Gunshot Types Based on Knowledge Distillation |
| title_short | A Fast Identification Method of Gunshot Types Based on Knowledge Distillation |
| title_sort | fast identification method of gunshot types based on knowledge distillation |
| topic | gunshots Log-Mel spectrum knowledge distillation |
| url | https://www.mdpi.com/2076-3417/12/11/5526 |
| work_keys_str_mv | AT jianli afastidentificationmethodofgunshottypesbasedonknowledgedistillation AT jinmingguo afastidentificationmethodofgunshottypesbasedonknowledgedistillation AT xiushansun afastidentificationmethodofgunshottypesbasedonknowledgedistillation AT chuankunli afastidentificationmethodofgunshottypesbasedonknowledgedistillation AT lingpengmeng afastidentificationmethodofgunshottypesbasedonknowledgedistillation AT jianli fastidentificationmethodofgunshottypesbasedonknowledgedistillation AT jinmingguo fastidentificationmethodofgunshottypesbasedonknowledgedistillation AT xiushansun fastidentificationmethodofgunshottypesbasedonknowledgedistillation AT chuankunli fastidentificationmethodofgunshottypesbasedonknowledgedistillation AT lingpengmeng fastidentificationmethodofgunshottypesbasedonknowledgedistillation |