|
|
import gradio as gr |
|
|
|
|
|
|
|
|
perception_papers = [ |
|
|
{ |
|
|
"title": "CoSDH: Communication-Efficient Collaborative Perception", |
|
|
"venue": "CVPR 2025", |
|
|
"description": "Novel approach for efficient collaborative perception using supply-demand awareness.", |
|
|
"link": "https://arxiv.org/abs/2503.03430" |
|
|
}, |
|
|
{ |
|
|
"title": "V2X-R: Cooperative LiDAR-4D Radar Fusion", |
|
|
"venue": "CVPR 2025", |
|
|
"description": "Cooperative fusion of LiDAR and 4D radar sensors for enhanced 3D object detection.", |
|
|
"link": "https://arxiv.org/abs/2411.08402" |
|
|
}, |
|
|
{ |
|
|
"title": "Where2comm: Efficient Collaborative Perception via Spatial Confidence Maps", |
|
|
"venue": "NeurIPS 2022", |
|
|
"description": "Groundbreaking work on efficient collaborative perception using spatial confidence maps.", |
|
|
"link": "https://openreview.net/forum?id=dLL4KXzKUpS" |
|
|
}, |
|
|
{ |
|
|
"title": "STAMP: Scalable Task-Agnostic Collaborative Perception", |
|
|
"venue": "ICLR 2025", |
|
|
"description": "Framework for scalable collaborative perception that is both task and model agnostic.", |
|
|
"link": "https://openreview.net/forum?id=8NdNniulYE" |
|
|
}, |
|
|
{ |
|
|
"title": "CoBEVFlow: Robust Asynchronous Collaborative 3D Detection", |
|
|
"venue": "NeurIPS 2023", |
|
|
"description": "Handles temporal asynchrony in collaborative perception using bird's eye view flow.", |
|
|
"link": "https://openreview.net/forum?id=UHIDdtxmVS" |
|
|
} |
|
|
] |
|
|
|
|
|
datasets_data = [ |
|
|
["DAIR-V2X", "2022", "Real-world", "V2I", "71K frames", "3D boxes, Infrastructure"], |
|
|
["V2V4Real", "2023", "Real-world", "V2V", "20K frames", "Real V2V, Highway"], |
|
|
["TUMTraf-V2X", "2024", "Real-world", "V2X", "2K sequences", "Dense labels, Urban"], |
|
|
["OPV2V", "2022", "Simulation", "V2V", "Large-scale", "CARLA, Multi-agent"], |
|
|
["V2X-Sim", "2021", "Simulation", "Multi", "Scalable", "Multi-agent, Collaborative"], |
|
|
["DOLPHINS", "2024", "Simulation", "UAV", "UAV swarms", "AirSim, Multi-UAV"] |
|
|
] |
|
|
|
|
|
def create_paper_card(paper): |
|
|
return f""" |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; margin: 10px 0; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<div style="background: #667eea; color: white; padding: 5px 10px; border-radius: 15px; display: inline-block; font-size: 0.8em; margin-bottom: 10px;"> |
|
|
{paper['venue']} |
|
|
</div> |
|
|
<h3 style="color: #333; margin: 10px 0;">{paper['title']}</h3> |
|
|
<p style="color: #666; line-height: 1.5; margin-bottom: 15px;">{paper['description']}</p> |
|
|
<a href="{paper['link']}" target="_blank" style="background: #667eea; color: white; padding: 8px 15px; border-radius: 5px; text-decoration: none; font-size: 0.9em;"> |
|
|
📄 Read Paper |
|
|
</a> |
|
|
</div> |
|
|
""" |
|
|
|
|
|
|
|
|
custom_css = """ |
|
|
.gradio-container { |
|
|
max-width: 1200px !important; |
|
|
} |
|
|
.main-header { |
|
|
text-align: center; |
|
|
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); |
|
|
color: white; |
|
|
padding: 40px 20px; |
|
|
border-radius: 15px; |
|
|
margin-bottom: 30px; |
|
|
} |
|
|
""" |
|
|
|
|
|
|
|
|
with gr.Blocks( |
|
|
title="🤖 Awesome Multi-Agent Collaborative Perception", |
|
|
theme=gr.themes.Soft(), |
|
|
css=custom_css |
|
|
) as demo: |
|
|
|
|
|
|
|
|
gr.HTML(""" |
|
|
<div class="main-header"> |
|
|
<h1 style="font-size: 2.5rem; margin-bottom: 10px;">🤖 Awesome Multi-Agent Collaborative Perception</h1> |
|
|
<p style="font-size: 1.2rem; opacity: 0.9;">Explore cutting-edge resources for Multi-Agent Collaborative Perception, Prediction, and Planning</p> |
|
|
<div style="display: flex; justify-content: center; gap: 30px; margin-top: 20px; flex-wrap: wrap;"> |
|
|
<div style="background: rgba(255,255,255,0.2); padding: 10px 20px; border-radius: 25px;"> |
|
|
<div style="font-size: 1.5rem; font-weight: bold;">200+</div> |
|
|
<div>Papers</div> |
|
|
</div> |
|
|
<div style="background: rgba(255,255,255,0.2); padding: 10px 20px; border-radius: 25px;"> |
|
|
<div style="font-size: 1.5rem; font-weight: bold;">25+</div> |
|
|
<div>Datasets</div> |
|
|
</div> |
|
|
<div style="background: rgba(255,255,255,0.2); padding: 10px 20px; border-radius: 25px;"> |
|
|
<div style="font-size: 1.5rem; font-weight: bold;">50+</div> |
|
|
<div>Code Repos</div> |
|
|
</div> |
|
|
<div style="background: rgba(255,255,255,0.2); padding: 10px 20px; border-radius: 25px;"> |
|
|
<div style="font-size: 1.5rem; font-weight: bold;">2025</div> |
|
|
<div>Updated</div> |
|
|
</div> |
|
|
</div> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
|
|
|
with gr.Tabs(): |
|
|
|
|
|
with gr.Tab("🔍 Perception"): |
|
|
gr.Markdown("## Multi-Agent Collaborative Perception Papers") |
|
|
gr.Markdown("*Latest research in collaborative sensing, 3D object detection, and V2X communication*") |
|
|
|
|
|
|
|
|
papers_html = "".join([create_paper_card(paper) for paper in perception_papers]) |
|
|
gr.HTML(papers_html) |
|
|
|
|
|
gr.Markdown(""" |
|
|
### 🔄 Key Communication Strategies: |
|
|
- **Early Fusion**: Raw sensor data sharing |
|
|
- **Late Fusion**: Detection-level information exchange |
|
|
- **Intermediate Fusion**: Feature-level collaboration |
|
|
- **Selective Communication**: Confidence-based data sharing |
|
|
""") |
|
|
|
|
|
with gr.Tab("📊 Datasets"): |
|
|
gr.Markdown("## Datasets & Benchmarks") |
|
|
gr.Markdown("*Comprehensive collection of real-world and simulation datasets*") |
|
|
|
|
|
gr.Dataframe( |
|
|
value=datasets_data, |
|
|
headers=["Dataset", "Year", "Type", "Agents", "Size", "Features"], |
|
|
datatype=["str", "str", "str", "str", "str", "str"], |
|
|
interactive=False |
|
|
) |
|
|
|
|
|
gr.Markdown(""" |
|
|
### 🌟 Notable Features: |
|
|
- **DAIR-V2X**: First real-world V2I collaborative perception dataset with infrastructure sensors |
|
|
- **V2V4Real**: Real vehicle-to-vehicle communication dataset collected on highways |
|
|
- **TUMTraf-V2X**: Dense annotations for urban collaborative perception scenarios |
|
|
- **OPV2V**: Large-scale simulation benchmark built on CARLA platform |
|
|
- **V2X-Sim**: Comprehensive multi-agent simulation with customizable scenarios |
|
|
""") |
|
|
|
|
|
with gr.Tab("📍 Tracking"): |
|
|
gr.Markdown("## Multi-Object Tracking & State Estimation") |
|
|
gr.Markdown("*Collaborative tracking across distributed agents with uncertainty quantification*") |
|
|
|
|
|
gr.HTML(""" |
|
|
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); gap: 20px; margin: 20px 0;"> |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<h3 style="color: #4ECDC4;">MOT-CUP</h3> |
|
|
<p>Multi-Object Tracking with Conformal Uncertainty Propagation</p> |
|
|
<a href="https://arxiv.org/abs/2303.14346" target="_blank" style="color: #667eea;">📄 Paper</a> |
|
|
</div> |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<h3 style="color: #4ECDC4;">DMSTrack</h3> |
|
|
<p>Probabilistic 3D Multi-Object Cooperative Tracking (ICRA 2024)</p> |
|
|
<a href="https://arxiv.org/abs/2309.14655" target="_blank" style="color: #667eea;">📄 Paper</a> |
|
|
</div> |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<h3 style="color: #4ECDC4;">CoDynTrust</h3> |
|
|
<p>Dynamic Feature Trust for Robust Asynchronous Collaborative Perception (ICRA 2025)</p> |
|
|
<a href="https://arxiv.org/abs/2502.08169" target="_blank" style="color: #667eea;">📄 Paper</a> |
|
|
</div> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
gr.Markdown(""" |
|
|
### 🎯 Key Challenges: |
|
|
- **Temporal Asynchrony**: Handling different sensor timestamps and communication delays |
|
|
- **Uncertainty Quantification**: Reliable confidence estimation across multiple agents |
|
|
- **Data Association**: Multi-agent correspondence and track management |
|
|
- **Scalability**: Maintaining performance with increasing number of agents |
|
|
""") |
|
|
|
|
|
with gr.Tab("🔮 Prediction"): |
|
|
gr.Markdown("## Trajectory Forecasting & Motion Prediction") |
|
|
gr.Markdown("*Cooperative prediction for autonomous systems and multi-agent coordination*") |
|
|
|
|
|
gr.HTML(""" |
|
|
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(300px, 1fr)); gap: 20px; margin: 20px 0;"> |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<h3 style="color: #45B7D1;">V2X-Graph</h3> |
|
|
<p>Learning Cooperative Trajectory Representations (NeurIPS 2024)</p> |
|
|
<a href="https://arxiv.org/abs/2311.00371" target="_blank" style="color: #667eea;">📄 Paper</a> |
|
|
</div> |
|
|
<div style="border: 1px solid #ddd; border-radius: 10px; padding: 20px; background: white; box-shadow: 0 2px 5px rgba(0,0,0,0.1);"> |
|
|
<h3 style="color: #45B7D1;">Co-MTP</h3> |
|
|
<p>Cooperative Multi-Temporal Prediction Framework (ICRA 2025)</p> |
|
|
<a href="https://arxiv.org/abs/2502.16589" target="_blank" style="color: #667eea;">📄 Paper</a> |
|
|
</div> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
gr.HTML(""" |
|
|
<div style="background: #f8f9fa; border-radius: 10px; padding: 20px; margin: 20px 0;"> |
|
|
<h3>🧠 Key Approaches:</h3> |
|
|
<ul style="line-height: 1.8;"> |
|
|
<li><strong>Graph Neural Networks</strong>: Modeling agent interactions and social behaviors</li> |
|
|
<li><strong>Transformer Architectures</strong>: Attention-based prediction with long-range dependencies</li> |
|
|
<li><strong>Multi-Modal Fusion</strong>: Combining LiDAR, camera, and communication data</li> |
|
|
<li><strong>Uncertainty Quantification</strong>: Reliable confidence estimation for safety-critical applications</li> |
|
|
</ul> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
with gr.Tab("⚙️ Methods"): |
|
|
gr.Markdown("## Methods & Techniques") |
|
|
gr.Markdown("*Core methodologies for communication, robustness, and learning in collaborative systems*") |
|
|
|
|
|
with gr.Row(): |
|
|
with gr.Column(): |
|
|
gr.Markdown(""" |
|
|
### 📡 Communication Strategies |
|
|
- **Bandwidth Optimization**: Compression and selective sharing |
|
|
- **Protocol Design**: V2V, V2I, V2X communication standards |
|
|
- **Network Topology**: Centralized vs. distributed architectures |
|
|
- **Quality of Service**: Latency and reliability management |
|
|
""") |
|
|
|
|
|
with gr.Column(): |
|
|
gr.Markdown(""" |
|
|
### 🛡️ Robustness Approaches |
|
|
- **Byzantine Fault Tolerance**: Handling adversarial agents |
|
|
- **Uncertainty Handling**: Robust fusion under noise |
|
|
- **Privacy Preservation**: Secure multi-party computation |
|
|
- **Malicious Agent Detection**: CP-Guard framework (AAAI 2025) |
|
|
""") |
|
|
|
|
|
gr.HTML(""" |
|
|
<div style="background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; border-radius: 10px; padding: 20px; margin: 20px 0;"> |
|
|
<h3>🧠 Learning Paradigms</h3> |
|
|
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(250px, 1fr)); gap: 15px; margin-top: 15px;"> |
|
|
<div>• <strong>Federated Learning</strong>: Distributed model training</div> |
|
|
<div>• <strong>Transfer Learning</strong>: Cross-domain adaptation</div> |
|
|
<div>• <strong>Meta-Learning</strong>: Quick adaptation to new scenarios</div> |
|
|
<div>• <strong>Heterogeneous Learning</strong>: HEAL framework (ICLR 2024)</div> |
|
|
</div> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
with gr.Tab("🏛️ Conferences"): |
|
|
gr.Markdown("## Top Venues & Publication Trends") |
|
|
gr.Markdown("*Premier conferences and emerging research directions in collaborative perception*") |
|
|
|
|
|
conference_data = [ |
|
|
["CVPR 2025", "5+", "End-to-end systems, robustness"], |
|
|
["ICLR 2025", "3+", "Learning representations, scalability"], |
|
|
["AAAI 2025", "4+", "AI applications, defense mechanisms"], |
|
|
["ICRA 2025", "6+", "Robotics applications, real-world deployment"], |
|
|
["NeurIPS 2024", "2+", "Theoretical foundations, novel architectures"] |
|
|
] |
|
|
|
|
|
gr.Dataframe( |
|
|
value=conference_data, |
|
|
headers=["Conference", "Papers", "Focus Areas"], |
|
|
datatype=["str", "str", "str"], |
|
|
interactive=False |
|
|
) |
|
|
|
|
|
gr.Markdown(""" |
|
|
### 📊 Research Trends (2024-2025): |
|
|
- **Communication Efficiency**: 40% increase in bandwidth-aware methods |
|
|
- **Robustness & Security**: Emerging focus on adversarial robustness (15+ papers) |
|
|
- **Real-World Deployment**: Growing emphasis on practical systems and field tests |
|
|
- **Heterogeneous Systems**: Multi-modal and multi-agent diversity becoming standard |
|
|
- **End-to-End Learning**: Integration of perception, prediction, and planning |
|
|
""") |
|
|
|
|
|
|
|
|
gr.HTML(""" |
|
|
<div style="text-align: center; margin-top: 40px; padding: 30px; background: #f8f9fa; border-radius: 10px;"> |
|
|
<h3>🤝 Contributing</h3> |
|
|
<p>We welcome contributions! Please submit papers, datasets, and code repositories via GitHub.</p> |
|
|
<div style="margin-top: 20px;"> |
|
|
<a href="https://github.com/your-username/awesome-multi-agent-collaborative-perception" target="_blank" |
|
|
style="background: #667eea; color: white; padding: 10px 20px; border-radius: 5px; text-decoration: none; margin: 5px;"> |
|
|
📚 GitHub Repository |
|
|
</a> |
|
|
<a href="https://huggingface.co/spaces/your-username/awesome-multi-agent-collaborative-perception" target="_blank" |
|
|
style="background: #ff6b6b; color: white; padding: 10px 20px; border-radius: 5px; text-decoration: none; margin: 5px;"> |
|
|
🤗 Hugging Face Space |
|
|
</a> |
|
|
</div> |
|
|
<p style="margin-top: 20px; color: #666;">Made with ❤️ for the Collaborative Perception Community</p> |
|
|
</div> |
|
|
""") |
|
|
|
|
|
if __name__ == "__main__": |
|
|
demo.launch() |
