Multi-label Image Recognition with Partial Labels
8 papers with code • 3 benchmarks • 2 datasets
Multi-label image recognition with partial labels (MLR-PL), in which some labels are known while others are unknown for each multi-label image, aims to train MLR models with partial labels to reduce the annotation cost. Since existing MLR datasets have complete labels, current works propose to randomly drop a certain proportion of positive and negative labels to create partially annotated datasets, and report the results on the known labels proportion of 10% to 90%.
Benchmarks
These leaderboards are used to track progress in Multi-label Image Recognition with Partial Labels
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
Visual Genome
PASCAL VOC 2007
Most implemented papers
Learning Semantic-Specific Graph Representation for Multi-Label Image Recognition
HCPLab-SYSU/SSGRL
•
•
ICCV 2019
Recognizing multiple labels of images is a practical and challenging task, and significant progress has been made by searching semantic-aware regions and modeling label dependency.
Learning Graph Convolutional Networks for Multi-Label Recognition and Applications
Kso-Ayaka/MS_Paperwithcode
•
•
IEEE Transactions on Pattern Analysis and Machine Intelligence 2021
The task of multi-label image recognition is to predict a set of object labels that present in an image.
Structured Semantic Transfer for Multi-Label Recognition with Partial Labels
To reduce the annotation cost, we propose a structured semantic transfer (SST) framework that enables training multi-label recognition models with partial labels, i. e., merely some labels are known while other labels are missing (also called unknown labels) per image.
Semantic-Aware Representation Blending for Multi-Label Image Recognition with Partial Labels
However, these algorithms depend on sufficient multi-label annotations to train the models, leading to poor performance especially with low known label proportion.
Heterogeneous Semantic Transfer for Multi-label Recognition with Partial Labels
Multi-label image recognition with partial labels (MLR-PL), in which some labels are known while others are unknown for each image, may greatly reduce the cost of annotation and thus facilitate large-scale MLR.
Dual-Perspective Semantic-Aware Representation Blending for Multi-Label Image Recognition with Partial Labels
Specifically, an instance-perspective representation blending (IPRB) module is designed to blend the representations of the known labels in an image with the representations of the corresponding unknown labels in another image to complement these unknown labels.
DualCoOp: Fast Adaptation to Multi-Label Recognition with Limited Annotations
Solving multi-label recognition (MLR) for images in the low-label regime is a challenging task with many real-world applications.
Texts as Images in Prompt Tuning for Multi-Label Image Recognition
Nonetheless, visual data (e. g., images) is by default prerequisite for learning prompts in existing methods.
