Unverified50% confidenceFactExact time
In cross-modal retrieval, text vectors and image vectors still cluster in separate regions within the shared space, causing overall similarity scores between text and images to be systematically lower—a phenomenon known as the Modality Gap.
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDomain Gap refers to the distribution difference between simulated images and real images.76% similarUnverifiedSemantic search for photos typically relies on image-text alignment models (such as CLIP-type multimodal embeddings) that map images and text into the same vector space.72% similarUnverifiedIn the field of face recognition, a low-discriminability feature space corresponds to a distribution where both intra-class variance and inter-class variance are small, causing Euclidean distance or cosine similarity metrics to fail.70% similarUnverified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68% similarUnverifiedStrong dependencies naturally exist between adjacent pixels in images, words in language, and phonemes in speech, making the conditional independence assumption invalid in most practical scenarios.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/985551API
curl https://kongchang.com/api/v1/knowledge/claims/985551MCP
get_claim(id=985551)