Linfeng-Tang / SeAFusion

The code of " Image fusion in the loop of high-level vision tasks: A semantic-aware real-time infrared and visible image fusion network"
MIT License
184 stars 34 forks source link

关于模型 #26

Open puyiwen opened 10 months ago

puyiwen commented 10 months ago

您好,我是这个方向的小白,我理解的是这个模型训出来是不是有两个.pth,一个是fusion image的模型,另外一个是分割模型?如果这样的话,那么做demo的时候是先要运行fusion image模型将RGB和IR做融合,生成融合图像。再将融合图像送入已经训练好的分割模型?期待您的回复~

Linfeng-Tang commented 10 months ago

是这样的

puyiwen commented 10 months ago

是这样的

好的谢谢您,您有计算过这两个模型的计算量和参数量吗?

Linfeng-Tang commented 10 months ago

论文里是没有计算的,你可以使用一些现成的工具进行统计。