Unverified50% confidenceFactExact time
Relative depth only guarantees that the near-far ordering and proportions between pixels are roughly reasonable, and does not promise absolute physical scale; metric depth requires output in real distances measured in meters, usually needing additional calibration or scale-annotated training data.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Perfect Depth Maps, Broken Stereo Video: Why?
rss10/15/2026
Related Entities
Related Claims
Unverified逐图 min/max 归一化会抹掉图像间的绝对强度差异，对依赖光谱绝对值区分地物的遥感分割任务可能丢失关键信息73% similarUnverifiedMonocular depth estimation requires only a single image, with a neural network inferring distance from visual cues such as brightness, perspective, and occlusion.67% similarUnverifiedLatent space is a core concept in deep generative models; during training, AI image models compress vast quantities of images into high-dimensional mathematical vectors, and the space formed by these vectors is the latent space.66% similarUnverified研究者已通过两点相关维数等方法实证测量了多种视觉模型的内在维度，训练良好的模型往往能将输入压缩到更低维流形上66% similarUnverified外观特征嵌入通过深度神经网络将目标图像块压缩映射为低维稠密向量，通常128维至512维65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1035738API
curl https://kongchang.com/api/v1/knowledge/claims/1035738MCP
get_claim(id=1035738)