Multilevel and Multiscale Feature Aggregation in Deep Networks for Facial Constitution Classification

<div>Visualization of the extracted features for the training data by t-SNE. The first row is for the 14<sup>th</sup> layer, and the second row is for the 24<sup>th</sup> layer. The left column is the original features, and the right column is the features after PCA.</div>

Computational and Mathematical Methods in Medicine

fig5

Figure 5

Figure 5: Multilevel and Multiscale Feature Aggregation in Deep Networks for Facial Constitution Classification 