a b/README.md
1
# DeepMesh
2
Motion tracking results of DeepMesh (the proposed method). 
3
Results are shown by overlaping meshes and images. Red contours are predicted results while green contours are ground truth
4
SAX | 2CH| 4CH 
5
--- | --- | --- 
6
<img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/sa_mid_pred.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/la_2ch_pred.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/la_4ch_pred.gif" width="120" height="120" /> 
7
8
Comparison methods
9
FFD | dDemons | 3D-UNet | MulViMotion|MeshMotion|DeepMesh(Proposed)
10
--- | --- | --- | ---| --- | ---
11
<img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/FFD_crop.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/dDemons_crop.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/3DUnet_crop.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/MulviMotion_crop.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/MeshMotion_crop.gif" width="120" height="120" /> | <img src="https://github.com/qmeng99/DeepMesh/blob/main/gifs/DeepMesh_crop.gif" width="120" height="120" />
12
13
14
# Network Architectures
15
16
**The network architecture of the Deformation network (H_D(·))**
17
<p align="center">
18
    <img src="https://github.com/qmeng99/DeepMesh/blob/main/network_architecture/DeformationNetwork.png" width="100%" height="100%">
19
</p>
20
21
22
**The network architecture of the Motion network (H_M(·))**
23
<p align="center">
24
    <img src="https://github.com/qmeng99/DeepMesh/blob/main/network_architecture/MotionNetwork.png" width="100%" height="100%">
25
</p>
26
27
28
**The network architecture for 2D feature learning in the Motion network**
29
<p align="center">
30
    <img src="https://github.com/qmeng99/DeepMesh/blob/main/network_architecture/2DFeatureLearning.png" width="100%" height="100%">
31
</p>
32
33
34
35
**The network architecture for 3D feature learning in the Motion network**
36
<p align="center">
37
    <img src="https://github.com/qmeng99/DeepMesh/blob/main/network_architecture/3DFeatureLearning.png" width="100%" height="100%">
38
</p>
39
40