Ensemble Convolution Neural Network for Robust Video Emotion Recognition Using Deep Semantics

Figure 2

Haar cascade features. (a) Line features. (b) Edge features. (c) Rectangle features.

(a)

(b)

(c)