Weakly-Supervised Object Localization
77 papers with code • 8 benchmarks • 3 datasets
Weakly supervised object localization (WSOL) learns to localize objects with only image-level labels, no object level labels (bonding boxes, etc.,) is needed. It is more attractive since image-level labels are much easier and cheaper to obtain.
Benchmarks
These leaderboards are used to track progress in Weakly-Supervised Object Localization
| Trend | Dataset | Best Model | Paper | Code | Compare |
|---|---|---|---|---|---|
|
|||||
|
|||||
|
|||||
|
|||||
|
|||||
|
|||||
|
|||||
|
Libraries
Use these libraries to find Weakly-Supervised Object Localization models and implementations
ImageNet
CUB-200-2011
Tiny ImageNet
Latest papers
Total Variation Optimization Layers for Computer Vision
raymondyeh07/tv_layers_for_cv
•
•
CVPR 2022
To study question (a), in this work, we propose total variation (TV) minimization as a layer for computer vision.
HINT: Hierarchical Neuron Concept Explainer
antonotnawang/hint
•
•
CVPR 2022
To this end, we propose HIerarchical Neuron concepT explainer (HINT) to effectively build bidirectional associations between neurons and hierarchical concepts in a low-cost and scalable manner.
Contrastive learning of Class-agnostic Activation Map for Weakly Supervised Object Localization and Semantic Segmentation
cvi-szu/ccam
•
•
While class activation map (CAM) generated by image classification network has been widely used for weakly supervised object localization (WSOL) and semantic segmentation (WSSS), such classifiers usually focus on discriminative object regions.
Weakly Supervised Object Localization as Domain Adaption
zh460045050/da-wsol_cvpr2022
•
•
CVPR 2022
To avoid this problem, this work provides a novel perspective that models WSOL as a domain adaption (DA) task, where the score estimator trained on the source/image domain is tested on the target/pixel domain to locate objects.
Self-Supervised Transformers for Unsupervised Object Discovery using Normalized Cut
YangtaoWANG95/TokenCut
•
•
CVPR 2022
For unsupervised saliency detection, we improve IoU for 4. 9%, 5. 2%, 12. 9% on ECSSD, DUTS, DUT-OMRON respectively compared to previous state of the art.
C2AM: Contrastive Learning of Class-Agnostic Activation Map for Weakly Supervised Object Localization and Semantic Segmentation
While class activation map (CAM) generated by image classification network has been widely used for weakly supervised object localization (WSOL) and semantic segmentation (WSSS), such classifiers usually focus on discriminative object regions.
Background-aware Classification Activation Map for Weakly Supervised Object Localization
timho102003/Background-Aware-CAM-WSOL
•
•
In our B-CAM, two image-level features, aggregated by pixel-level features of potential background and object locations, are used to purify the object feature from the object-related background and to represent the feature of the pure-background sample, respectively.
Group-Wise Learning for Weakly Supervised Semantic Segmentation
Lixy1997/Group-WSSS
•
•
journal 2021
The framework explicitly encodes semantic dependencies in a group of images to discover rich semantic context for estimating more reliable pseudo ground-truths, which are subsequently employed to train more effective segmentation models.
Background Activation Suppression for Weakly Supervised Object Localization
wpy1999/bas
•
•
CVPR 2022
Existing FPM-based methods use cross-entropy (CE) to evaluate the foreground prediction map and to guide the learning of generator.
TDAM: Top-Down Attention Module for Contextually Guided Feature Selection in CNNs
shantanuj/tdam_top_down_attention_module
•
•
Attention modules for Convolutional Neural Networks (CNNs) are an effective method to enhance performance on multiple computer-vision tasks.
