Unverified50% confidenceFactExact time
YOLOv3采用多尺度特征金字塔（FPN）
1
Sources
50%
Confidence
Long-term
Relevance
7/25/2026
First Seen
Sources
Related Claims
UnverifiedYOLOv3引入多尺度预测和残差网络；YOLOv4集成了CSPNet和PANet。78% similarUnverifiedYOLO V3采用Darknet-53骨干网络并借鉴特征金字塔（FPN）思想，引入三尺度检测头77% similarUnverifiedYOLO11采用C3k2模块替代传统C2f，并使用改进的PANet路径聚合网络进行特征融合70% similarUnverified相比YOLOv8，YOLOv11在模型架构上引入了C3k2模块和SPPF改进，在保持检测精度的同时降低了参数量和计算量70% similarUnverifiedYOLOv8 模型家族从 YOLOv8n（约 3.2M 参数）到 YOLOv8x（约 68M 参数）覆盖从树莓派到高性能 GPU 的硬件谱系66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/615487API
curl https://kongchang.com/api/v1/knowledge/claims/615487MCP
get_claim(id=615487)