TADP: Task-Aware Deformable Prediction for Single-Stage 3D Object Detection

TADP: Task-Aware Deformable Prediction for Single-Stage 3D Object Detection

|Authors: Su Wang, Yaochen Li, Min Yang, Jiaohao Nie, Chang Liu, Yuehu Liu

|Venue: arXiv preprint | Year: 2026

|arXiv: 2608.27282v1

Abstract

Most single-stage 3D object detectors complete different tasks with the same extracted features. Nevertheless, it is impossible to project features into a common space that is adaptive for all the tasks. We present a novel task-aware deformable prediction (TADP) method for single-stage 3D object detection to solve this problem. Firstly, a triple feature refinement aggregation module is designed to extract three-level features adaptively. Additionally, we design the multi-scale feature aggregation block to fuse multi-scale features in a scale-aware manner. Finally, the prediction of each task is deformed with the designed plug-and-play task-aware deformation head. It can percept the emphasis and interaction of each task. We also designed three different deformation modules. The experimental results demonstrate that the proposed deformation head shows good results on other detection methods. The experimental results on the KITTI dataset demonstrate that the car mAP is 80.91%, surpassing many state-of-the-art methods on the KITTI benchmark.

Key Contributions

• Most single-stage 3D object detectors complete different tasks with the same extracted features
• Nevertheless, it is impossible to project features into a common space that is adaptive for all t…
• We present a novel task-aware deformable prediction (TADP) method for single-stage 3D object dete…
• Firstly, a triple feature refinement aggregation module is designed to extract three-level featur…
• Additionally, we design the multi-scale feature aggregation block to fuse multi-scale features in…

Topics

• vision
• benchmark

Code & Data

No code repository linked in paper metadata.

BibTeX

@article{Wang2026_260827282v1,
  title     = {TADP: Task-Aware Deformable Prediction for Single-Stage 3D Object Detection},
  author    = {Su Wang and Yaochen Li and Min Yang and Jiaohao Nie and Chang Liu and Yuehu Liu},
  year      = {2026},
  eprint    = {2608.27282v1},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url       = {https://arxiv.org/abs/2608.27282v1}
}