search

Linfeng-Tang / SuperFusion

This is official Pytorch implementation of "SuperFusion: A Versatile Image Registration and Fusion Network with Semantic Awareness"
MIT License
62 stars 9 forks source link

SuperFusion #8

Open mm111222 opened 11 months ago

mm111222 commented 11 months ago

Can you give the results of training with registration and fusion? Best_RegFusion.pth

Linfeng-Tang commented 11 months ago

Can you describe your problem in further detail? We provide model files that are trained with registration and fusion together.

mm111222 commented 11 months ago

I want to register and fuse the trained model without subsequent image fusion and segmentation training

Linfeng-Tang commented 11 months ago

The RoadScene.pth is only trained with register and fusion.

mm111222 commented 10 months ago

Whether the CrossRAFT model in your comparison experiment was retrained by yourself or in the original paper Can you provide the trained model parameters?

Linfeng-Tang commented 10 months ago

The CrossRAFT model is provided by the original paper.

mm111222 commented 10 months ago

The test after my own deformation is generally lower than the index in your paper, and I suspect that it is caused by the different degree of my deformation。Can you send me the 361 MSRS anamorphic images that I used to test the indicators?

mm111222 commented 10 months ago

Fig. 9. How is it drawn and what is the meaning?

Linfeng-Tang commented 10 months ago

A point (x; y) on the curve denotes that there are 100  x percent of image pairs that have metric values no more than y.