We present SegLLM, a novel multi-round interactive segmentation model that leverages conversational memory of both visual and textual outputs to reason over previously segmented objects and past ...
Abstract: Existing surface defect semantic segmentation methods are limited by costly annotated data and are unable to cope with new or rare defect types. Zero-shot learning offers a new possibility ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback