Factorized convolutional networks: unsupervised fine-tuning for image clustering

March 2018

Factorized convolutional networks: unsupervised fine-tuning for image clustering

Authors:

Liang-Yan Gui, Liangke Gui, Yuxiong Wang, Louis-Philippe Morency, and Jose M. F. Moura

Abstract:

Deep convolutional neural networks (CNNs) have recognized promise as universal representations for various image recognition tasks. One of their properties is the ability to transfer knowledge from a large annotated source dataset (e.g., ImageNet) to a (typically smaller) target dataset. This is usually accomplished through supervised fine-tuning on labeled new target data. In this work, we address "unsupervised fine-tuning" that transfers a pre-trained network to target tasks with unlabeled data such as image clustering tasks. To this end, we introduce group-sparse non-negative matrix factorization (GSNMF), a variant of NMF, to identify a rich set of high-level latent variables that are informative on the target task. The resulting "factorized convolutional network" (FCN) can itself be seen as a feed-forward model that combines CNN and two-layer structured NMF. We empirically validate our approach and demonstrate state-of-the-art image clustering performance on challenging scene (MIT-67) and fine-grained (Birds-200, Flowers-102) benchmarks. We further show that, when used as
@conference{Gui-2018-122552,
author = {Liang-Yan Gui And Liangke Gui And Yuxiong Wang And Louis-Philippe Morency And Jose M. F. Moura},
title = {Factorized convolutional networks: unsupervised fine-tuning for image clustering},
booktitle = {Proceedings of IEEE Winter Conference on Applications of Computer Vision (WACV '18)},
year = {2018},
month = {March},
pages = {1205 - 1214},
}
Copyright notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.