Continually Expanding Retrieval-Augmented Segmentation Models without Retraining

A novel training-free framework, kNN-CLIP, that continually expands the vocabulary of image segmentation models by leveraging a dynamic retrieval database, without the need for retraining.