基于对比学习的声源定位引导视听分割模型

黄文湖,赵邢,谢亮,梁浩然,梁荣华

Contrastive learning-based sound source localization-guided audio-visual segmentation model

Wenhu HUANG,Xing ZHAO,Liang XIE,Haoran LIANG,Ronghua LIANG

表 5 ACT激活对实验结果的影响

Tab.5 Effect of ACT activation on experimental results

激活函数 图像编码器 S4 MS3 $ {M_{\text{J}}} $/% $ {M_{\text{F}}} $/% $ {M_{\text{J}}} $/% $ {M_{\text{F}}} $/% ACT ResNet-50 78.68 88.0 59.50 69.8 PVT v2 84.43 91.6 65.16 75.6 Sigmoid ResNet-50 78.19 87.8 55.40 66.1 PVT v2 84.15 91.3 61.69 72.9 — ResNet-50 77.76 87.2 58.76 69.3 PVT v2 84.33 91.5 63.30 74.3