Abstract:Multi-label image classification aims to predict all possible labels in an image. It is usually formulated as a partial-label learning problem, given the fact that it could be expensive in practice to annotate all labels in every training image. Existing works on partial-label learning focus on the case where each training image is annotated with only a subset of its labels. A special case is to annotate only one positive label in each training image. To further relieve the annotation burden and enhance the performance of the classifier, this paper proposes a new partial-label setting in which only a subset of the training images are labeled, each with only one positive label, while the rest of the training images remain unlabeled. To handle this new setting, we propose an end-to-end deep network, PLMCL (Partial Label Momentum Curriculum Learning), that can learn to produce confident pseudo labels for both partially-labeled and unlabeled training images. The novel momentum-based law updates soft pseudo labels on each training image with the consideration of the updating velocity of pseudo labels, which help avoid trapping to low-confidence local minimum, especially at the early stage of training in lack of both observed labels and confidence on pseudo labels. In addition, we present a confidence-aware scheduler to adaptively perform easy-to-hard learning for different labels. Extensive experiments demonstrate that our proposed PLMCL outperforms many state-of-the-art multi-label classification methods under various partial-label settings on three different datasets.

Multi-instance multi-label new label learning

Dual Enhancement for Multi-Label Learning with Missing Labels

Multi-Label Learning with Emerging New Labels

Multi-Instance Multi-Label Learning with Application to Scene Classification

Learning from Multi-Dimensional Partial Labels.

Collaboration based multi-modal multi-label learning

Labeling Complicated Objects: Multi-View Multi-Instance Multi-Label Learning

Joint multi-label multi-instance learning for image classification

Discover Multiple Novel Labels in Multi-Instance Multi-Label Learning.

Two-Stage Label Embedding Via Neural Factorization Machine for Multi-Label Classification

A New multi-instance multi-label learning approach for image and text classification

Dual Set Multi-Label Learning

Multi-Instance Multi-Label Learning with Weak Label.

Multi-Label Learning with Block Diagonal Labels

PLMCL: Partial-Label Momentum Curriculum Learning for Multi-Label Image Classification

Weakly-Supervised Multi-view Multi-instance Multi-label Learning

Rethinking Modal-oriented Label Correlations for Multi-modal Multi-label Learning

Learning Disentangled Label Representations for Multi-label Classification

M3LA: A Novel Approach Based on Encoder-Decoder with Attention Framework for Multi-modal Multi-label Learning

Dynamic Ensemble Learning for Multi-label Classification

MULTI-LABEL IMAGE RECOGNITION WITH JOINT CLASS-AWARE MAP DISENTANGLING AND LABEL CORRELATION EMBEDDING