Skip to content
/ AudioMLProject2 Public

Speaker Detection: We have one target speaker whom we want to detect. The set of impostors is open, i.e., we have no prior information on the test speakers. An important part of the speaker detection pipeline is voice activity detection to filter out segments of the signal that do not contain speech.

7 stars 4 forks Branches Tags Activity
Star
Notifications

netankit/AudioMLProject2

BranchesTags

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

33 Commits

AdaBoostClassifier.py

AdaBoostClassifier.py

 
 

FrameLevelAnnotation.py

FrameLevelAnnotation.py

 
 

FrameLevelAnnotation_FileWise.py

FrameLevelAnnotation_FileWise.py

 
 

FrameLevelAnnotation_speakerubm.py

FrameLevelAnnotation_speakerubm.py

 
 

IVector.py

IVector.py

 
 

KmeansUBM.py

KmeansUBM.py

 
 

MAP_Enrollment_final.py

MAP_Enrollment_final.py

 
 

README.md

README.md

 
 

ivecter_score_cosine_distance.py

ivecter_score_cosine_distance.py

 
 

ivecter_score_lda.py

ivecter_score_lda.py

 
 

ivecter_score_plda.py

ivecter_score_plda.py

 
 

machine_training.py

machine_training.py

 
 

plot_test.py

plot_test.py

 
 

prediction_score.py

prediction_score.py

 
 

project_ivectors.py

project_ivectors.py

 
 

remove_zeros.py

remove_zeros.py

 
 

remove_zeros_filewise.py

remove_zeros_filewise.py

 
 

scoring_GMM_UBM_final.py

scoring_GMM_UBM_final.py

 
 

stack_train_test_vectors.py

stack_train_test_vectors.py

 
 

train_test_dataset.py

train_test_dataset.py

 
 

vad_mixing_pathfixed.py

vad_mixing_pathfixed.py

 
 

wccn_test.py

wccn_test.py

 
 

Repository files navigation

Speaker Detection – from GMMs to i-vectors

Goals

  • Mix corpora
  • Create a GMM based speaker detection system with MAP adaption
  • Create an i-vector based speaker detection system
  • Evaluate
  • Present your results in form of a report

About

Speaker Detection: We have one target speaker whom we want to detect. The set of impostors is open, i.e., we have no prior information on the test speakers. An important part of the speaker detection pipeline is voice activity detection to filter out segments of the signal that do not contain speech.

Resources

Readme
Activity

Stars

7 stars

Watchers

3 watching

Forks

4 forks
Report repository

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •  

Languages