Update README.md
Browse files
README.md
CHANGED
|
@@ -5,4 +5,26 @@ tags:
|
|
| 5 |
- biology
|
| 6 |
- cancer
|
| 7 |
- microscopy
|
| 8 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
- biology
|
| 6 |
- cancer
|
| 7 |
- microscopy
|
| 8 |
+
license: other
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
See github here for usage:
|
| 12 |
+
https://github.com/ReyaLab/AI_Colony_Counter
|
| 13 |
+
|
| 14 |
+
https://reya-lab.org/
|
| 15 |
+
|
| 16 |
+
Custom fine-tuned version of NVIDIA's segformer model for colony slides in microscopy.
|
| 17 |
+
|
| 18 |
+
Xie, E., Wang, W., Yu, Z., Anandkumar, A., Alvarez, J. M., & Luo, P. (2021). SegFormer: Simple and efficient design for semantic segmentation with transformers. arXiv preprint arXiv:2105.15203. https://arxiv.org/abs/2105.15203
|
| 19 |
+
|
| 20 |
+
The Segformer-Cityscapes model was changed to a ternary classifier and fine-tuned on custom training data, where colonies and necrosis were made as separate masks and then merged with different grayscale values.
|
| 21 |
+

|
| 22 |
+

|
| 23 |
+

|
| 24 |
+
|
| 25 |
+
After training, it was able to correctly identify organoids and necrosis.
|
| 26 |
+

|
| 27 |
+

|
| 28 |
+
|
| 29 |
+
The python program (see linked GitHub) then uses the masks to annotate the images and provide statistics about the colonies.
|
| 30 |
+

|