Image Segmentation
Transformers
English
fabda commited on
Commit
652ec03
·
1 Parent(s): 2f91a5a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -7
README.md CHANGED
@@ -2,20 +2,37 @@
2
  license: gpl-3.0
3
  language:
4
  - en
 
5
  ---
6
 
7
- ### 3DL_NuCount model
8
 
9
- Original Paper: _M1BP is an essential transcriptional activator of oxidative metabolism during Drosophila flight muscle development
10
- Gabriela Poliacikova, Marine Barthez, Thomas Rival, Aïcha Aouane, Nuno Miguel Luis, Fabrice Richard, Fabrice Daian, Nicolas Brouilly, Frank Schnorrer, Corinne Maurel-Zaffran, Yacine Graba and Andrew J. Saurin**_
11
 
12
 
13
  #### Model description
14
 
15
- __3DL_NuCount__ model has been designed by fine tuning a pretrained Stardist3D model using a home made dataset in order to
16
 
 
17
 
18
- #### References
19
- - Stardist Project:[Github](https://github.com/stardist/stardist)
20
- - Stardist Paper : [ArXiv](https://arxiv.org/abs/1908.03636)
 
 
 
 
 
 
 
 
21
 
 
 
 
 
 
 
 
 
 
2
  license: gpl-3.0
3
  language:
4
  - en
5
+ pipeline_tag: image-segmentation
6
  ---
7
 
8
+ # 3DL_NuCount model
9
 
10
+ __Paper__: M1BP is an essential transcriptional activator of oxidative metabolism during Drosophila flight muscle development. _Gabriela Poliacikova, Marine Barthez, Thomas Rival, Aïcha Aouane, Nuno Miguel Luis, Fabrice Richard, Fabrice Daian, Nicolas Brouilly, Frank Schnorrer, Corinne Maurel-Zaffran, Yacine Graba and Andrew J. Saurin**_
 
11
 
12
 
13
  #### Model description
14
 
15
+ __3DL_NuCount__ model has been designed by fine tuning a pretrained Stardist3D model [1,2] using a home made dataset [3] in order to assess the number of cells present in a given 3D image stack acquired using an optical microscope.
16
 
17
+ #### Stardist Training parameters
18
 
19
+ - patch size: (48,96,96)
20
+ - batch size: 32
21
+ - epochs : 100
22
+ - data augmentation : flip/rotation/intensity
23
+ - image normalization: normalize channel independantly
24
+ - anisotropy: empirical
25
+ - rays : 96
26
+
27
+ #### Training dataset parameters
28
+ - tile size : (4,63,128,128)
29
+ - split : Train 0.8 / Val 0.2
30
 
31
+ #### Inference
32
+ - patch size: (784,784,:)
33
+ - image size: (2048,2048,:)
34
+
35
+ #### References
36
+ - [1] Stardist Project:[Github](https://github.com/stardist/stardist)
37
+ - [2] Stardist Paper : [ArXiv](https://arxiv.org/abs/1908.03636)
38
+ - [3] Training Data : [Zenodo](https://)