Tianheng Cheng, Lin Song, Yixiao Ge, Wenyu Liu, Xinggang Wang, Ying Shan
The You Only Look Once (YOLO) series of detectors have established themselves as efficient and practical tools. However, their reliance on predefined and trained object categories limits their applicability in open scenarios. Addressing this limitation, we introduce YOLO-World, an innovative approach that enhances YOLO with open-vocabulary detection capabilities through vision-language modeling and pre-training on large-scale datasets. Specifically, we propose a new Re-parameterizable Vision-Language Path Aggregation Network (RepVL-PAN) and region-text contrastive loss to facilitate the interaction between visual and linguistic information. Our method excels in detecting a wide range of objects in a zero-shot manner with high efficiency. On the challenging LVIS dataset, YOLO-World achieves 35.4 AP with 52.0 FPS on V100, which outperforms many state-of-the-art methods in terms of both accuracy and speed. Furthermore, the fine-tuned YOLO-World achieves remarkable performance on several downstream tasks, including object detection and open-vocabulary instance segmentation.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Object Detection | LVIS v1.0 minival | AP | 35.4 | YOLO-World-L |
| Object Detection | MSCOCO | AP | 45.1 | YOLO-World-L(without COCO data) |
| 3D | LVIS v1.0 minival | AP | 35.4 | YOLO-World-L |
| 3D | MSCOCO | AP | 45.1 | YOLO-World-L(without COCO data) |
| 2D Classification | LVIS v1.0 minival | AP | 35.4 | YOLO-World-L |
| 2D Classification | MSCOCO | AP | 45.1 | YOLO-World-L(without COCO data) |
| 2D Object Detection | LVIS v1.0 minival | AP | 35.4 | YOLO-World-L |
| 2D Object Detection | MSCOCO | AP | 45.1 | YOLO-World-L(without COCO data) |
| 16k | LVIS v1.0 minival | AP | 35.4 | YOLO-World-L |
| 16k | MSCOCO | AP | 45.1 | YOLO-World-L(without COCO data) |