Fig 1.
Inter-floor noise recording scene.
Fig 2.
Sound level for 24 hours.
Fig 3.
Recorded audio length and main sources.
Fig 4.
(a) Footsteps, (b) Dragging furniture, (c) Vacuum cleaner, (d) Hammering, (e) Instant impact and (f) PA system.
Table 1.
Inter-floor dataset configuration.
Fig 5.
Model size and accuracy.
Table 2.
Classification results (5-fold cross validation).
Fig 6.
Confusion matrix of ResNet.
Table 3.
Comparison to other datasets.
Table 4.
Comparison to the state-of-the-art models.