No description, website, or topics provided.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
doc
matlab
python
README.md
Results.txt

README.md

MCE18-THUEE

THUEE system for The 1st Multi-target speaker detection and identification Challenge Evaluation
http://www.mce2018.org/

environment : anaconda3, python3, require sklearn.

  1. prepare data, make directory ./data and ./temp
    put MCE18 offical uncompressed data on "./data/",
    there are "bl_matching_dev.csv, trn_blacklist.csv, trn_background.csv, dev_blacklist.csv, dev_background.csv, tst_evaluation.csv"
    PS: open the website (http://www.mce2018.org/) for data requirement.

There are two systems:
A. length-normalization + LPLDA + Cosine + score-normalization
B. lenght-normalization + LDA + PLDA (MotPLDA) + score-normalization

System A:
A. run ./python/mce18_lplda.py
If we use LDA (line 158, "lda = LinearDiscriminantAnalysis(n_components=500)"), the results will be
Dev set score using train set :
Top S detector EER is 2.76%
Top 1 detector EER is 9.20% (Total confusion error is 325)

If we use LPLDA (line 159, "lda = LPLDA.LocalPairwiseLinearDiscriminantAnalysis(n_components=500)"), the results will be
Dev set score using train set :
Top S detector EER is 2.41%
Top 1 detector EER is 8.10% (Total confusion error is 291)

System B:
A. run ./python/mce18_plda_preprocess.py
It will generate "./temp/mce18.mat"

Option 1: B1. run ./matlab/gplda_demo.m for PLDA
The script will read "./temp/mce18.mat", and it will generate "./temp/mce18_result.mat"

C1. run ./python/mce18_plda_eval.py
The script will read "./temp/mce18.mat", and the results are
Dev set score using train set :
Top S detector EER is 2.77%
Top 1 detector EER is 9.04% (Total confusion error is 322)

Option 2: B2. run ./matlab/moplda_demo.m for MoPLDA
The script will read "./temp/mce18.mat", and it will also generate "./temp/mce18_result.mat"

C2. run ./python/mce18_plda_eval.py
The script will read "./temp/mce18.mat", and the results are
Dev set score using train set :
Top S detector EER is 3.83%
Top 1 detector EER is 7.50% (Total confusion error is 255)

[1] Suwon Shon, Najim Dehak, Douglas Reynolds, and James Glass, “Mce 2018: The 1st multi-target speaker detection and identification challenge evaluation (mce) plan, dataset and baseline system,” in ArXiv e-prints arXiv:1807.06663, 2018.

[2] L. He, X. Chen, C. Xu, J. Liu, and M. T. Johnson, “Local pairwise linear discriminant analysis for speaker verification,” IEEE Signal Processing Letters, 2018.

[3] L. He, X. Chen, C. Xu, and J. Liu, “Multiobjective Optimization Training of PLDA for Speaker Verification,” ArXiv e-prints arXiv:1808.08344, Aug. 2018.

He Liang, heliang@mail.tsinghua.edu.cn Sep. 26, 2018