Prompt-Guided Transformers for End-to-End Open-Vocabulary Object Detection
This addresses the problem of efficient and accurate detection of both base and novel object classes for computer vision applications, representing a strong incremental improvement.
The paper tackles open-vocabulary object detection by proposing Prompt-OVD, a framework that uses CLIP class embeddings as prompts to guide a Transformer decoder, achieving 21.2 times faster inference than OV-DETR and higher APs than comparable two-stage methods.
Prompt-OVD is an efficient and effective framework for open-vocabulary object detection that utilizes class embeddings from CLIP as prompts, guiding the Transformer decoder to detect objects in both base and novel classes. Additionally, our novel RoI-based masked attention and RoI pruning techniques help leverage the zero-shot classification ability of the Vision Transformer-based CLIP, resulting in improved detection performance at minimal computational cost. Our experiments on the OV-COCO and OVLVIS datasets demonstrate that Prompt-OVD achieves an impressive 21.2 times faster inference speed than the first end-to-end open-vocabulary detection method (OV-DETR), while also achieving higher APs than four two-stage-based methods operating within similar inference time ranges. Code will be made available soon.