Abstract:This paper considers deep visual recognition on long-tailed data. Under the long-tailed distribution, a small portion of the classes (head classes) occupy most training samples and the most classes (tail classes) only occupy relatively few samples. We observe that such long-tailed distribution significantly distorts the deeply-learned feature space, which consequentially compromises the deep visual recognition. Specifically, during training, each head class is prone to a relatively wide spatial distribution in the deep feature space, while each tail class is prone to a relatively small spatial distribution. In another word, the tail classes usually have much smaller spatial distribution than the head classes, distorting the overall feature space. In response, we propose to explicitly inflate the distribution of each tail class in the deep feature space, so that the tail classes will have comparable distribution range as the head classes. To this end, we replace each tail feature vector with a set of feature vectors on the fly. These feature vectors follow a probabilistic distribution learned from the head classes and yield a “feature cloud” surrounding the original tail feature. We show that the feature cloud effectively transfers the within-class diversity from the head classes onto the tail classes, maintaining an effect of probabilistic feature augmentation. An important advantage of the proposed feature cloud is that it is capable to bring general improvement to long-tailed visual recognition on two fundamental tasks, i.e., deep classification and deep representation learning, in spite of the significant differences between them. Extensive experiments on both deep metric learning benchmarks and deep image classification benchmarks validate the effectiveness of the proposed feature cloud.

Instance-Specific Semantic Augmentation for Long-Tailed Image Classification

SAFA: Sample-Adaptive Feature Augmentation for Long-Tailed Image Classification

ProAug: Prototype-Based Augmentation for Long-Tailed Image Classification.

Feature Space Augmentation for Long-Tailed Data

BSDA in Visual Recognition: Balanced Semantic Data Augmentation for Long-Tailed Data.

ECS-SC: Long-tailed classification via data augmentation based on easily confused sample selection and combination

Label-Specific Feature Augmentation for Long-Tailed Multi-Label Text Classification

Data-Centric Long-Tailed Image Recognition

Imagine by Reasoning: A Reasoning-Based Implicit Semantic Data Augmentation for Long-Tailed Classification

MetaSAug: Meta Semantic Augmentation for Long-Tailed Visual Recognition

Text-guided Fourier Augmentation for long-tailed recognition

Kill Two Birds with One Stone: Rethinking Data Augmentation for Deep Long-tailed Learning

Attentive Feature Augmentation for Long-Tailed Visual Recognition

DDFA: a displacement and diffusion-based feature augmentation method for imbalanced image recognition

SGIA: Enhancing Fine-Grained Visual Classification with Sequence Generative Image Augmentation

A Dual Progressive Strategy for Long-Tailed Visual Recognition

Feature Cloud: Improving Deep Visual Recognition with Probabilistic Feature Augmentation

Class Activation Maps-based Feature Augmentation for long-tailed classification

Location-Based Scene Reconstruction for Long-Tail Recognition

CUDA: Curriculum of Data Augmentation for Long-Tailed Recognition

Improving Long-Tailed Classification from Instance Level