Deep unfolding network improves unsupervised feature selection in images
Learning to Select Features with Manifold Deep Unfolding PCA
Computer Vision and Pattern RecognitionMachine Learning
Summary
Finding the most important features in data without labels is hard and usually requires manual tuning or lots of training examples. The authors created a special network that turns a classic data analysis method into a trainable model that learns parameters automatically. They also add a technique to make the learning more stable and a second version to pick features that better separate data groups. Their tests show these approaches work better than previous methods at choosing features and helping group data points correctly.
What this means in practice
- •For image processing teams: Improve unsupervised feature selection in image datasets to enhance downstream clustering and recognition tasks without requiring extensive manual parameter tuning.
- •For data engineers: Automate feature selection in large unlabeled datasets using a robust trainable network that adapts parameters for better data representation quality.
Authors
Jingjing Liu, Pei Tai, Xianchao Xiu, Wanquan Liu
Abstract
Unsupervised feature selection (UFS) is of great significance in image processing and pattern recognition. Nevertheless, existing model-based methods often rely on manually tuned hyperparameters, while deep learning-based methods generally require massive training data. To address these issues, this paper proposes an efficient manifold deep unfolding feature selection network (MDUFS-Net). Technically, we first incorporate the $\ell_{2,0}$-norm into the classic principal component analysis (PCA) framework, then unfold the alternating direction method of multipliers (ADMM)-based iterative algorithm into an end-to-end trainable network, thereby enabling automatic parameter learning. Moreover, a manifold optimization strategy is developed to improve the numerical robustness. On this basis, we further extend MDUFS-Net into an enhanced variant called MDUFS-Net+ by introducing the $\ell_{0}$-norm for more discriminative feature representations. Extensive experiments validate that our proposed MDUFS-Net and MDUFS-Net+ outperform state-of-the-art methods in the UFS task, with promising performance in clustering. Our code is available at https://github.com/xianchaoxiu/MDUFS.