DeCo-DETR: Decoupled Cognition DETR for efficient Open-Vocabulary Object Detection
This addresses efficiency and generalization issues in open-vocabulary object detection for practical deployment, representing an incremental improvement.
The paper tackles the problem of high computational overhead and trade-offs in open-vocabulary object detection by proposing DeCo-DETR, a vision-centric framework that uses a hierarchical semantic prototype space and decoupled training. It achieves competitive zero-shot detection performance with significantly improved inference efficiency.
Open-vocabulary Object Detection (OVOD) enables models to recognize objects beyond predefined categories, but existing approaches remain limited in practical deployment. On the one hand, multimodal designs often incur substantial computational overhead due to their reliance on text encoders at inference time. On the other hand, tightly coupled training objectives introduce a trade-off between closed-set detection accuracy and open-world generalization. Thus, we propose Decoupled Cognition DETR (DeCo-DETR), a vision-centric framework that addresses these challenges through a unified decoupling paradigm. Instead of depending on online text encoding, DeCo-DETR constructs a hierarchical semantic prototype space from region-level descriptions generated by pre-trained LVLMs and aligned via CLIP, enabling efficient and reusable semantic representation. Building upon this representation, the framework further disentangles semantic reasoning from localization through a decoupled training strategy, which separates alignment and detection into parallel optimization streams. Extensive experiments on standard OVOD benchmarks demonstrate that DeCo-DETR achieves competitive zero-shot detection performance while significantly improving inference efficiency. These results highlight the effectiveness of decoupling semantic cognition from detection, offering a practical direction for scalable OVOD systems.