Skip to main content
. 2023 Jan 30;89:101754. doi: 10.1016/j.pmcj.2023.101754

Fig. 4.

Fig. 4

Memory footprint of (a) the considered shallow classifiers based on different input sizes (i.e., percentage of PCA explained variance), and (b) the fine-tuned deep learning models. Please, consider that in (a) the overall memory footprint is given by taking into account also the size of the deep embedding models to extract the input features from the raw audio sample.