asvspoof2019 - GitLite

KO | EN

gitlite — search

#javascript #python #react #machine-learning #rust #deep-learning #linux #nodejs #awesome #vue #awesome-list #pytorch #ios #open-source

asvspoof2019

(★ 104)

Our submission to the ASVspoof 2019: Automatic Speaker Verification Spoofing and Countermeasures Challenge

파일 탐색기

최종 버전 다운로드 (.zip)

- - .nfs000000000c80c5400002541b
- - .DS_Store
  - cqcc.m
  - D18_1000001.wav
  - DEMO.m
  - README.txt
- - .DS_Store
  - compute_llk.m
  - lgmmprob.m
  - logsumexp.m
- - .DS_Store
  - deltas.m
  - extract_lfcc.m
- - BOSARIS_License.txt
  - compute_det_curve.m
  - compute_eer.m
  - demo_tDCF.m
  - evaluate_tDCF_asvspoof19.m
  - LICENSE
  - obtain_asv_error_rates.m
  - README.txt
- ASVspoof2019_baseline_CM.m
- LICENSE
- Readme.txt
- - cqcc.pth
- - mfcc.pth
  - mfcc_delta_delta2.pth
  - mfcc_logical.pth
- - spec.pth
  - spec_logical.pth
  - spec_physical2.pth
  - spect_finetune.pth
  - spect_physical.pth
- .nfsAD1C
- fuse_result.txt
- FusionFunction.ipynb
- LICENSE
- merge_result.csv
- README.txt
- Untitled.ipynb
.gitignore
cqcc_extraction.m
data_utils.py
download_data.bash
eval_metrics.py
evaluate_tDCF_asvspoof19.py
fuse_result.py
LICENSE
model_main.py
models.py
README.md
requirements.txt

# 설치 가이드

1. 코드 내려받기

git clone https://github.com/nesl/asvspoof2019

깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.

cd asvspoof2019

방금 내려받은 프로젝트 폴더 안으로 이동합니다.

2. Python

쉬움 추천

사전 준비물

Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
Python 3 설치 시 'Add Python to PATH' 옵션을 꼭 체크하세요 (Windows).

pip install -r requirements.txt

requirements.txt 등에 명시된 파이썬 라이브러리를 설치합니다.

jupyter notebook

브라우저에서 노트북(.ipynb) 파일들을 열람하고 실행할 수 있는 Jupyter 화면을 켭니다.

✅ 에러 메시지 없이 실행되고 터미널에 안내 문구가 출력되면 정상입니다.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Deep Residual Neural Networks for Audio Spoofing Detection

This repo has an implementation for our paper **Deep Residual Neural Networks for Audio Spoofing Detection**, this is also describes the solution of team **UCLANESL** in the [ASVSpoof 2019 competition](https://www.asvspoof.org/).

## Dataset

The ASVSpoof2019 dataset can be downloaded from the following link:

[ASVSpoof2019 dataset](https://datashare.is.ed.ac.uk/handle/10283/3336)

### Training models
```
python model_main.py --num_epochs=100 --track=[logical/physical] --features=[spect/mfcc/cqcc]   --lr=0.00005
```

Please note that the CQCC features are computing using the Matlab code in [cqcc_extraction.m](./cqcc_extraction.m), so you need to run this file to generate cache files of CQCC featurs before attempting to traiin or evaluate models with CQCC features.

#### To perform fusion of multiple results files
```
 python fuse_result.py --input FILE1 FILE2 FILE3 --output=RESULTS_FILE
```

### Evaluating Models

Run the model on the evaluation dataset to generate a prediction file.
```
python model_main.py --eval  --eval_output=RESULTS_FILE --model_path=CHECKPOINT_FILE
```

Then compute the evaluation scores using on the development dataset

```
python evaluate_tDCF_asvspoof19.py RESULTS_FILE PATH_TO__asv_dev.txt 
```