음성학적 과학수사 용도의 달팽이관 모사 스펙트럼을 이용하는 딥러닝에 기반한 대화자 식별 알고리즘
Speaker Identification Algorithm Based on the Deep Learning for Phonetics Forensic Purposes using a Cochlear Simulation Spectrum
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
제한된 증거를 바탕으로 하는 과학수사 분야에서는 최소한의 데이터로부터 최선의 결과를 얻을 수있는 알고리즘의 개발이 필요하다. 본 연구에서는 평균 0.556초의 짧은 음성을 이용하는 화자인식 딥러닝알고리즘을 개발하였다. 개발에 사용된 달팽이관 모사 스펙트럼은 음성 신호의 기존 파형을 최대한 유지하면서 인간이 청각 신호를 받아드리는 전처리 과정을 모사하여 딥러닝 네트워크가 분류에 집중하여 학습할수 있도록 하였다. 그 결과, 2차원 합성곱 신경망에 71명의 화자 중 1명의 음성을 무작위로 제시하였을 때교차검증 평균 정확도 96.2%로 1명의 화자를 식별하였다. 본 연구에서 수집한 데이터베이스는 거짓말탐지자극검사를 활용하여 추후 거짓말 탐지 연구에도 활용할 수 있을 뿐 아니라, 개발한 알고리즘은 전과자 데이터베이스를 활용하여 용의자를 특정하는 데에 활용할 수 있을 것으로 기대된다.
Abstract
In the field of forensic investigation based on limited evidence, it is essential to develop an algorithm that can derive the best performance on a given minimal data. In this study, the authors developed a deep learning algorithm capable of speaker recognition using a short speech of an average of 0.556 seconds. The cochleagram used in the development preserves the existing waveform of the speech signal as much as possible while simulating the preprocessing process in humans auditory system. Therefore, the network can focus on classification task while learning. As a result, when the utterance of one of 71 speakers was input to the 2D convolutional neural network, the average speaker identification accuracy of 10 fold cross-validation was 96.2%. The collected database can be used for later lie detection research because it reflecting the stimulus test used in the lie detection investigation. And expect the developed algorithm can be used to identify new suspects using the ex-convict database.
참고문헌 (19)
- 조호대, "보이스피싱 발생 및 대응방안", 한국콘텐츠학회 논문지, 12(7), pp.176~182 (2012), 한국콘텐츠학회[DOI]
- DA Reynolds, "Speaker identification and verification using Gaussian mixture speaker models", Speech communication, 17(1-2), pp.91~108 (1995)
- WM Campbell, "Support vector machines using GMM supervectors for speaker verification", IEEE signal processing letters, 13(5), pp.308~311 (2006)
- B Logan, "Mel frequency cepstral coefficients for music modeling", ISMIR, 270, pp.1~11 (2000)
- D Garcia-Romero, "Analysis of i-vector length normalization in speaker recognition systems", Twelfth annual conference of the international speech communication association, pp.249~252 (2011)
- A Krizhevsky, "Imagenet classification with deep convolutional neural networks", Communications of the ACM, 60(6), pp.84~90 (2017)
- S Hochreiter, "Long short-term memory", Neural computation, 9(8), pp.1735~1780 (1997)
- HS Lee, "Speaker verification using kernel-based binary classifiers with binary operation derived features", 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp.1660~1664 (2014), IEEE
- NF Dronkers, "Encyclopedia of neuroscience" (2009), Elsevier
- 이경태, "달팽이관 필터뱅크 및 ZCPA 청각모델에 기반한 두 귀의 시간차 추정기법", 한국소음진동공학회논문집, 29(6), pp.722~734 (2019), 한국소음진동공학회[DOI]
- L Saxe, "Scientific Validity of Polygraph Testing-A Review and Evaluation-A Technical Memorandum", Polygraph, 12(3), pp.198~319 (1983)
- D Griffin, "Signal estimation from modified short-time Fourier transform", IEEE Transactions on Acoustics, Speech, and Signal Processing, 32(2), pp.236~243 (1984)
- RV Sharan, "Cochleagram image feature for improved robustness in sound recognition", 2015 IEEE International Conference on Digital Signal Processing (DSP), pp.441~444 (2015), IEEE
- R Kohavi, "A study of cross-validation and bootstrap for accuracy estimation and model selection", Ijcai, 14(2), pp.1137~1145 (1995)
- P Cremonesi, "Performance of recommender algorithms on top-n recommendation tasks", Proceedings of the fourth ACM conference on Recommender systems, pp.39~46 (2010)
- S Nakagawa, "Speaker identification and verification by combining MFCC and phase information", IEEE transactions on audio, speech, and language processing, 20(4), pp.1085~1095 (2011)
- V Tiwari, "Speaker identification using multi-modal i-vector approach for varying length speech in voice interactive systems", Cognitive Systems Research, 57, pp.66~77 (2019)
- D Hendrycks, "A baseline for detecting misclassified and out-of-distribution examples in neural networks", arXiv preprint arXiv:1610.02136, ~ (2016)
- S Liang, "Enhancing the reliability of out-of- distribution image detection in neural networks", arXiv preprint arXiv:1706.02690, ~ (2017)