Efficient and accurate joint representation of a collection of images, that belong to the same class, is a major research challenge for practical image set classification. Existing methods either make prior assumptions about the data structure, or perform heavy computations to learn structure from the data itself. We propose a Deep Extreme Learning Machine (DELM) for efficient learning of the nonlinear structure of image sets without making any assumption about the underlying data. The DELM generalizes very well based on a limited number of training samples. We learn a domain specific DELM model in an unsupervised fashion and then adapt it to learn class specific representations. Extensive experiments on a broad range of public datasets for image set classification (Honda/UCSD, CMU Mobo, YouTube Celebrities, Celebrity-1000, ETH-80) show that the proposed DELM consistently outperforms state-ofthe- art image set classification methods both in terms of speed and accuracy.
View on arXiv