Voiceprint_Recognition

Just for DCASE 2019 and for studying AI

data

Task name : Acoustic Scene Classfication
Download : You can download the datasets from here datasets , and the code for datasets are "wzzg"

data preprocess

We use skip_frames in our preprocess stage, the skip is 2. In our future work, we will use sparse_frames and skip_frames together.

features

mfcc174

model

Densenet

cost function

CrossEntropy

Operation process

read_file.py --> file_shuffle.py --> same_step.py --> feature_extract.py --> DenseNet.py

reference link

https://github.com/qiuqiangkong/audioset_classification