[20180511] Deep Covariance Descriptors for Facial Expression Recognition

[unizard]

BMVC2018 submission

URL: https://arxiv.org/pdf/1805.03869.pdf Keyword: Deep Covariance Descriptor Interest: 2

Summary

By performing the classification of the facial expressions using Gaussian kernel on SPD manifold, we show that the covariance descriptors computed on DCNN features are more efficient than the standard classification with fully connected layers and softmax.

기존 방법과 달리 Covariance descriotor 기반의 DCNN 피쳐를 사용했더니, 감정인식성능이 올랐다.

[unizard]

Can we improve IR performance using "DELF + Covariance descriptor"? Traditionally, covariance is robust in deformation. Especially, in terms of ratio...

[unizard]

Landmarks are clustered in five group. Is it done, automatically?