Research map / Multi-agent systems and games / Human-AI and goal-conditioned
IFTR: An Instance-Level Fusion Transformer for Visual Collaborative Perception
Shaohong Wang, Lu Bin, Xinyu Xiao, Zhiyu Xiang, Hangguan Shan, Eryun Liu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add instance feature aggregation (IFA) with BEV queries to late fusion (ablation) | BEVFormer-S with late fusion · OPV2V | AP@70: 0.5192 -> 0.6487 | Improved |
| add cross-domain query adaptation (CDQA) (ablation) | IFTR with IFA · OPV2V | AP@70: 0.6487 -> 0.6604 | Improved |
| add cross-domain query adaptation (CDQA) (ablation) | IFTR with IFA · DAIR-V2X | AP@70: 0.0574 -> 0.0665 | Improved |
| share only foreground features (MASK) (ablation) | IFTR · OPV2V | AP@70 and communication: AP@70 0.6604 -> 0.6300; Comm 24.23 -> 20.61 | Mixed |
| share only foreground features (MASK) (ablation) | IFTR · DAIR-V2X | AP@70 and communication: AP@70 0.0665 -> 0.0618; Comm 21.91 -> 18.23 | Mixed |
| use cone encoding instead of no positional encoding in CDQA (ablation) | CDQA · DAIR-V2X | AP@70: 0.0610 -> 0.0665 | Improved |
| use cone encoding instead of learned positional encoding (ablation) | CDQA · OPV2V | AP@70: 0.6588 -> 0.6604 | Improved |
| increase number of collaborating agents | IFTR · OPV2V | 3D detection AP | Improved |
| replace BEV-level fusion methods with IFTR under localization noise | V2XSet and DAIR-V2X · localization noise 0-0.6m | AP | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Where2comm — Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps (NeurIPS 2022) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- STAMP: Scalable Task- And Model-agnostic Collaborative Perception (ICLR 2025 · similarity 0.90)
- Privacy-Preserving V2X Collaborative Perception Integrating Unknown Collaborators (AAAI 2025 · similarity 0.89)
- An Extensible Framework for Open Heterogeneous Collaborative Perception (ICLR 2024 · similarity 0.89)
- One is Plenty: A Polymorphic Feature Interpreter for Immutable Heterogeneous Collaborative Perception (CVPR 2025 · similarity 0.88)
- IS-Fusion: Instance-Scene Collaborative Fusion for Multimodal 3D Object Detection (CVPR 2024 · similarity 0.88)
- AGC-Drive: A Large-Scale Dataset for Real-World Aerial-Ground Collaboration in Driving Scenarios (NeurIPS 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
