Deep Imbalanced Learning for Face Recognition and Attribute Prediction

Abstract

Data for face analysis often exhibit highly-skewed class distribution, i.e., most data belong to a few majority classes, while the minority classes only contain a scarce amount of instances. To mitigate this issue, contemporary deep learning methods typically follow classic strategies such as class re-sampling or cost-sensitive training. In this paper, we conduct extensive and systematic experiments to validate the effectiveness of these classic schemes for representation learning on class-imbalanced data. We further demonstrate that more discriminative deep representation can be learned by enforcing a deep network to maintain inter-cluster margins both within and between classes. This tight constraint effectively reduces the class imbalance inherent in the local data neighborhood, thus carving much more balanced class boundaries locally. We show that it is easy to deploy angular margins between the cluster distributions on a hypersphere manifold. Such learned Cluster-based Large Margin Local Embedding (CLMLE), when combined with a simple k-nearest cluster algorithm, shows significant improvements in accuracy over existing methods on both face recognition and face attribute prediction tasks that exhibit imbalanced class distribution.

arXiv comment14 pages, 10 figures, 8 tables. Accepted to TPAMI

References in corpus (8)

• SMOTE: Synthetic Minority Over-sampling Technique
• Deep Learning Face Representation by Joint Identification-Verification
• Learning Face Representation from Scratch
• Learning to Reweight Examples for Robust Deep Learning
• Targeting Ultimate Accuracy: Face Recognition via Deep Embedding
• Rethinking Feature Discrimination and Polymerization for Large-scale Recognition
• One-shot Face Recognition by Promoting Underrepresented Classes
• Feature Transfer Learning for Deep Face Recognition with Under-Represented Data

Cited by in corpus (10)

• Feature Fusion from Head to Tail for Long-Tailed Visual Recognition
• Dynamic Curriculum Learning for Imbalanced Data Classification
• The Devil of Face Recognition is in the Noise
• Instance-level Facial Attributes Transfer with Geometry-Aware Flow
• Meta-GCN: A Dynamically Weighted Loss Minimization Method for Dealing with the Data Imbalance in Graph Neural Networks
• Consensus-Driven Propagation in Massive Unlabeled Data for Face Recognition
• Deep Imbalanced Attribute Classification using Visual Attention Aggregation
• Consistent Batch Normalization for Weighted Loss in Imbalanced-Data Environment
• Improvement of Batch Normalization in Imbalanced Data
• On Class Imbalance and Background Filtering in Visual Relationship Detection