The dataset contains 164 recordings of cat sounds (1,323 seconds) and 113 recordings of dog sounds (598 seconds). The length of the recordings is variable.
Retrieve the resource:
Retrieve the dataset:
Select an entry from the dataset:
Visualize the waveform of the Audio object:
Visualize the spectrum:
Visualize the spectrogram:
Compute the number of examples per class:
Inspect the sample rate and channel count of the Audio objects:
Plot the histogram of the durations of the Audio objects:
By computing the cumulative duration of the labels, it can be seen that the dataset is not balanced:
The average recording duration for the two labels is different as well:
Create a balanced training dataset.
Make sure that the class total duration of the selected samples is similar:
Join the two classes and shuffle the elements:
Create an unbalanced testing dataset using the remaining data:
"Audio Cats and Dogs"
from the Wolfram Data Repository
Data Resource History