
The author proposes a training-free approach, FineR, that leverages Large Language Models (LLMs) to reason about fine-grained category names from visual descriptions, eliminating the need for expert annotations.


coremsg

democratizing-fine-grained-visual-recognition-with-large-language-models


Democratizing Fine-grained Visual Recognition with Large Language Models


title_rewrite


専門家のアノテーション不要で、大規模言語モデルを活用して細かい視覚カテゴリを理解する方法を提案。


大規模言語モデルを活用した細かい視覚認識の民主化