基于多视图跨模态特征融合的图像描述生成

张乃洲,赵云超,曹薇,张啸剑

Image captioning generation based on multiple-view cross-modal feature fusion

Naizhou ZHANG,Yunchao ZHAO,Wei CAO,Xiaojian ZHANG

表 5 MVCMFAF模型与其他模型在计算量、参数量和推理时间方面的比较

Tab.5 Comparison of computational complexity, parameter quantity and inference time between MVCMFAF model and other model

模型 FLOPs/109 Np/MB t/ms Xmodal-Ctx[16] 127.614 35.439 137.107 DIFNet[8] 137.412 28.395 98.244 PureT[21] 882.301 224.201 238.937 MVCMFAF (本文) 137.461 175.769 446.157