Commit
•
99a0f33
1
Parent(s):
dade98a
Update README.md
Browse files
README.md
CHANGED
@@ -10,7 +10,7 @@ This model card describes the model associated with the manuscript "Uncertainty-
|
|
10 |
- **Model type:** Deep convolutional neural network image classifier
|
11 |
- **Language(s):** English
|
12 |
- **License:** GPL-3.0
|
13 |
-
- **Model Description:** This is a model that can classify H&E-stained pathologic images of non-small cell lung cancer into adenocarcinoma or squamous cell carcinoma and provide an estimate of classification uncertainty. It is an [Xception](https://arxiv.org/abs/1610.02357) model with two dropout-enabled hidden
|
14 |
- **Image processing:** This model expects images of H&E-stained pathology slides at 299 x 299 px and 302 x 302 μm resolution. Images should be stain-normalized using a modified Reinhard normalizer ("Reinhard-Fast") available [here](https://github.com/jamesdolezal/slideflow/blob/master/slideflow/norm/tensorflow/reinhard.py). The stain normalizer should be fit using the `target_means` and `target_stds` listed in the model `params.json` file. Images should be should be standardized with `tf.image.per_image_standardization()`.
|
15 |
- **Resources for more information:** [GitHub Repository](https://github.com/jamesdolezal/biscuit), [Paper](https://www.nature.com/articles/s41467-022-34025-x)
|
16 |
- **Cite as:**
|
|
|
10 |
- **Model type:** Deep convolutional neural network image classifier
|
11 |
- **Language(s):** English
|
12 |
- **License:** GPL-3.0
|
13 |
+
- **Model Description:** This is a model that can classify H&E-stained pathologic images of non-small cell lung cancer into adenocarcinoma or squamous cell carcinoma and provide an estimate of classification uncertainty. It is an [Xception](https://arxiv.org/abs/1610.02357) model with two dropout-enabled hidden layers enabled during both training and inference. During inference, a given image is passed through the network 30 times, resulting in a distribution of predictions. The mean of this distribution is the final prediction, and the standard deviation is the uncertainty.
|
14 |
- **Image processing:** This model expects images of H&E-stained pathology slides at 299 x 299 px and 302 x 302 μm resolution. Images should be stain-normalized using a modified Reinhard normalizer ("Reinhard-Fast") available [here](https://github.com/jamesdolezal/slideflow/blob/master/slideflow/norm/tensorflow/reinhard.py). The stain normalizer should be fit using the `target_means` and `target_stds` listed in the model `params.json` file. Images should be should be standardized with `tf.image.per_image_standardization()`.
|
15 |
- **Resources for more information:** [GitHub Repository](https://github.com/jamesdolezal/biscuit), [Paper](https://www.nature.com/articles/s41467-022-34025-x)
|
16 |
- **Cite as:**
|