/JuvenileSong_pipeline

Primary LanguagePython

JuvenileSong_pipeline

Preprocessing Steps

python main.py INPUT_folder
  1. dat2wav matlab/electro_gui
  2. combine: gather around 25 bouts in one file raw data
  3. filter: bandfilter applied filtered
  4. denoise denoised

Data Labeling with DAS

conda activate DAS
das gui

4 TCN block used. feedback loop used.

Postprocessing Steps: UMAP

python umap_yirong.py INPUT_folder ‘wav’

use trained model (with data from 7 days) to predict data on bigger time-range. UMAP visualize predict annotation, to check how's it performance. https://github.com/AFurryBear/syllable_modify