Unverified50% confidenceBenchmarkExact time
The directional uniformity measure method has the best overall performance on synthetic data and is more robust to noise and non-uniform density.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/19/2026
First Seen
Valid until: 1/17/2027
Sources
Related Entities
Related Claims
Unverified扩散模型在噪声水平高的早期去噪步骤对精度相对不敏感，在噪声水平低的后期精细化步骤对数值扰动更敏感71% similarUnverified网格对齐时，均值对高斯噪声最优，中位数对椒盐噪声更鲁棒，众数最能保留原始颜色意图69% similarUnverifiedThe directional uniformity measure method is sensitive to neighborhood size and noise, and its statistical efficiency may decline in ultra-high dimensions.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1065025API
curl https://kongchang.com/api/v1/knowledge/claims/1065025MCP
get_claim(id=1065025)