search

ahof1704 / VocalMat

Analysis of ultrasonic vocalizations from mice using computer vision and machine learning
https://www.dietrich-lab.org/vocalmat
Apache License 2.0
22 stars 10 forks source link

All_axes folder not appearing #12

Closed cmlenell1 closed 2 years ago

cmlenell1 commented 2 years ago

Hello, I installed the program and ran on several files but for some reason the "all_axes" folder is missing from the outputs. Everything else seems to be working fine. Do you have any suggestions?

ahof1704 commented 2 years ago

Hi,

The all_axes folder should be created when you call for the vocalmat_classifier function as seen here. https://github.com/ahof1704/VocalMat/blob/9384fabfc1fbd9bc0ef8ca460b652e72c5b6819f/vocalmat_classifier/vocalmat_classifier.m#L214

Do you get any error messages in your prompt? Which OS are you using?

cmlenell1 commented 2 years ago

Hi, I don't have any related errors. I attached screenshots of the matlab code running the file. matlab1 matlab2

ahof1704 commented 2 years ago

Have you changed the "save_plot_spectrograms" to "1" (which means you want the spectrograms to be saved). Note that it is "0" by default.

https://github.com/ahof1704/VocalMat/blob/9384fabfc1fbd9bc0ef8ca460b652e72c5b6819f/VocalMat.m#L30

cmlenell1 commented 2 years ago

That fixed it. Thanks so much!

As a side question, do you recommend manually reviewing usv categories that have a certain probability to avoid misclassifying ~14% of the usvs?

ahof1704 commented 2 years ago

I would recommend having a general inspection of the detected and classified USVs, just to make sure things are working properly (ie, evaluate if you had too much noise in your recordings and if that is interfering with detection and classification of the USVs). Once you verify that things are working as expected, you can double-check the USVs classified as one of the less common categories (see Fig.4c of the paper) or with probabilities lower than 85%.

Good luck!