BenchmarkJensen-Shannon Distance感知分布度量
JOD
A perceptual distribution metric that measures the overall difference between generated results and ground truth at the level of feature distributions, and is sensitive to blur and artifacts.
Timeline (last 90 days)
Oct 19
On the novel-view reenactment task, VDAs on the BecomingLit dataset reduces facial LPIPS by 29% and improves JOD by 66% compared with the strongest single-image 3D avatar method.
Unverified50%