SOTAVerified

Visual Grounding

Visual Grounding (VG) aims to locate the most relevant object or region in an image, based on a natural language query. The query can be a phrase, a sentence, or even a multi-round dialogue. There are three main challenges in VG:

  • What is the main focus in a query?
  • How to understand an image?
  • How to locate an object?

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 526–550 of 571 papers

TitleStatusHype
Leveraging Past References for Robust Language Grounding—0
Countering Language Drift via Visual Grounding—0
Language learning using Speech to Image retrieval—0
Differentiable Disentanglement Filter: an Application Agnostic Core Concept Discovery Probe—0
Multimodal Unified Attention Networks for Vision-and-Language Interactions—0
Differentiable Disentanglement Filter: an Application Agnostic Core Concept Discovery Probe—0
Transfer Learning from Audio-Visual Grounding to Speech Recognition—0
Joint Visual Grounding with Language Scene Graphs—0
Visually Grounded Neural Syntax Acquisition—0
Learning to Compose and Reason with Language Tree Structures for Visual Grounding—0
On the Contributions of Visual and Textual Supervision in Low-Resource Semantic Speech Retrieval—0
Semantic query-by-example speech search using visual groundingCode0
Modularized Textual Grounding for Counterfactual ResilienceCode0
VQD: Visual Query Detection in Natural Scenes—0
Revisiting Visual Grounding—0
Learning semantic sentence representations from visually grounded language without lexical knowledgeCode0
Align2Ground: Weakly Supervised Phrase Grounding Guided by Image-Caption Alignment—0
Dual Attention Networks for Visual Reference Resolution in Visual DialogCode0
You Only Look & Listen Once: Towards Fast and Accurate Visual GroundingCode0
Taking a HINT: Leveraging Explanations to Make Vision and Language Models More Grounded—0
Learning to Assemble Neural Module Tree Networks for Visual Grounding—0
Multi-task Learning of Hierarchical Vision-Language Representation—0
Being data-driven is not enough: Revisiting interactive instruction giving as a challenge for NLG—0
Overcoming Language Priors in Visual Question Answering with Adversarial Regularization—0
Beyond task success: A closer look at jointly learning to see, ask, and GuessWhatCode0
Show:102550
← PrevPage 22 of 23Next →
All datasetsRefCOCO testARefCOCO+ test BRefCoCo val

Benchmark Results

#ModelMetricClaimedVerifiedStatus
1Florence-2-large-ftAccuracy (%)95.3—Unverified
2mPLUG-2Accuracy (%)92.8—Unverified
3X2-VLM (large)Accuracy (%)92.1—Unverified
4XFM (base)Accuracy (%)90.4—Unverified
5X2-VLM (base)Accuracy (%)90.3—Unverified
6X-VLM (base)Accuracy (%)89—Unverified
7HYDRAIoU61.7—Unverified
8HYDRAIoU61.1—Unverified
#ModelMetricClaimedVerifiedStatus
1Florence-2-large-ftAccuracy (%)92—Unverified
2mPLUG-2Accuracy (%)86.05—Unverified
3X2-VLM (large)Accuracy (%)81.8—Unverified
4XFM (base)Accuracy (%)79.8—Unverified
5X2-VLM (base)Accuracy (%)78.4—Unverified
6X-VLM (base)Accuracy (%)76.91—Unverified
#ModelMetricClaimedVerifiedStatus
1Florence-2-large-ftAccuracy (%)93.4—Unverified
2mPLUG-2Accuracy (%)90.33—Unverified
3X2-VLM (large)Accuracy (%)87.6—Unverified
4XFM (base)Accuracy (%)86.1—Unverified
5X2-VLM (base)Accuracy (%)85.2—Unverified
6X-VLM (base)Accuracy (%)84.51—Unverified