Audio Segmentation

The goal of audio segmentation is to split an un-iterrupted audio signal into homogeneous segments.

audioSegment.py 是pyAudioAnalysislibrary裡的檔案。

圖0 Schematic view of segmentation methods in audioSegmentation.py file

圖0呈現audio segmentation的分類示意圖。

audio segmentation方法都是joint segmentation-classification，意思是指在把音頻切成小片段時，並把小片段做標註(e.g. silence vs non-silence)。

There are two type of audio segmentation$\colon$ supervised audio segmentation, unsupervised audio segmentation.

In supervised Segmentation, there are three methods $\colon$

1. fixed-size-segment
2. KNN
3. HMM

In unsupervised Segmentation, there are three applications $\colon$

1. silence removal
2. speaker diarization
3. music audio thumbnailing

[0] https://github.com/tyiannak/pyAudioAnalysis/wiki/5.-Segmentation

[1] Giannakopoulos, Theodoros, and Sergios Petridis. "Fisher linear semi-discriminant analysis for speaker diarization." IEEE Transactions on Audio, Speech, and Language Processing 20.7 (2012): 1913-1922.