SEA: Understanding Sketch Abstraction Efficiency via Element-Level Commonsense for Human-AI Sketching
SEA is introduced, a reference-free metric that measures how economically a sketch preserves class-defining visual elements while maintaining semantic recognizability and enables fine-grained evaluation of sketch understanding in vision-language models (VLMs).
Jiho Park, Sieun Choi, Jaeyoon Seo et al.
· 0 citations