This document summarizes a research paper that proposes a new deep convolutional neural network (DCNN) ensemble classifier for facial expression recognition in uncontrolled environments. The key aspects of the proposed method are: (1) Formulating the process of finding optimal ensemble weights as a stochastic optimization problem solved using simulated annealing. This allows the weights to minimize generalized classification error. (2) Creating diverse DCNN ensemble members by combining different face representations and bagging, which increases ensemble diversity. Experiments on three facial expression datasets show the proposed DCNN ensemble achieves state-of-the-art performance.

Combining Deep Convolutional Neural Networks With Stochastic Ensemble Weight Optimization For Facial Expression Recognition in The Wild PDF

The list of books you might like