Unverified50% confidenceFactExact time
PSNR measures pixel-level error between a reconstructed image and a reference image in decibels on a logarithmic scale, where a 1 dB improvement typically corresponds to a perceptible improvement in detail.
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLPIPS (Learned Perceptual Image Patch Similarity) has a far higher correlation with human subjective evaluation than traditional PSNR and SSIM metrics.69% similarVerifiedFID measures the distributional distance between generated images and real images in feature space; a lower value indicates higher generation quality.65% similarUnverified现有基准测试如WebSRC、Design2Code主要测量像素还原度，采用SSIM或CLIP相似度作为量化指标，但高还原度与高设计质量并不等价63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/975907API
curl https://kongchang.com/api/v1/knowledge/claims/975907MCP
get_claim(id=975907)