Unverified50% confidenceFactExact time
现代卷积神经网络的感受野随层级加深而递增，功能上与人类视觉皮层V1→V2→V4→IT区的加工流水线高度吻合
1
Sources
50%
Confidence
Long-term
Relevance
7/21/2026
First Seen
Sources
Related Claims
Unverified视觉皮层中从高级脑区回传到低级脑区的反馈连接数量往往超过前馈连接74% similarUnverifiedConvolutional Neural Networks (CNNs), inspired by the structure of the visual cortex, extract local features by sliding convolutional kernels across images, with the parameter sharing mechanism significantly reducing computational load.73% similarUnverifiedIn image recognition, shallow neurons respond to low-level features such as edges and textures, while deep neurons respond to overall concepts.71% similarUnverifiedSpatial cognition research indicates that the hippocampus and entorhinal cortex of the human brain form a powerful spatial navigation and spatial memory system.68% similarUnverified长期多语言使用者的前额叶皮层、扣带回及基底节区域往往表现出更高的灰质密度和更强的功能性连接67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/575733API
curl https://kongchang.com/api/v1/knowledge/claims/575733MCP
get_claim(id=575733)