[78ef36]: / docs / _sources / overview.rst.txt

Download this file

66 lines (37 with data), 4.2 kB

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Overview
========
Slideflow provides tools for easily building and testing a variety of deep learning models for digital pathology.
This section provides a high-level overview of the most common application: building and testing a weakly supervised predictive model. Slideflow supports many other tasks, including :ref:`multiple-instance learning (MIL) <mil>`, :ref:`self-supervised learning (SSL) <simclr_ssl>`, :ref:`generative adversarial networks (GANs) <stylegan>`, :ref:`tissue <segmentation>` and :ref:`cell <cellseg>` segmentation, and :ref:`deployment & visualization <studio>`, which are discussed in subsequent sections.
.. figure:: overview.png
*High-level overview of model building.*
The pipeline for a deep learning classification experiment is separated into three phases.
1) **Tile extraction** - annotate slides with regions of interest (ROIs) [*optional*] and extract image tiles from whole-slide images.
2) **Model training** - determine model parameters, train a model, and evaluate the model on a held-out test set.
3) **Explainability** - generate predictive heatmaps and analyze learned image features.
|
A brief introduction to the steps needed to execute a basic experiment is provided below. Each process will be described in more detail in the following sections.
Step 1: Prepare a dataset
*************************
- **Extract tiles**. :ref:`Tiles are extracted <filtering>` from slides at a given magnification size in microns (or a magnification layer, such as "10x"), and saved at a given resolution in pixels. The optimal extraction size in both microns and pixels will depend on your dataset and model architecture. Poor quality tiles - including background tiles or tiles with high whitespace content - can be discarded with quality control methods. Tiles will be stored as TFRecords, a binary file format used to improve dataset reading performance during training. Each slide will have its own TFRecord file containing its extracted tiles.
- **Set aside final evaluation set**. :ref:`Split the dataset <datasets_and_validation>` into a training/validation set and held-out test set.
- **Determing validation plan**. By default, three-fold cross-validation will be performed during training. Many other validation strategies are also supported (:ref:`validation_planning`).
Step 2: Train a model
*********************
- **Choose model type**. Choose the endpoint (e.g. classification, regression, time-to-event) and type of model (tile-based or multiple-instance learning).
- **Set hyperparameters**. Choose a model architecture (e.g. InceptionV3, VGG16, ResNet, etc.) and a set of hyperparameters (e.g. batch size, learning rate, etc.). This can be done manually, or :ref:`hyperparameters can be optimized <hyperparameter_optimization>` via grid search or Bayesian optimization.
- **Initiate training**. :ref:`Train your model <training>`, taking note of training and validation performance (e.g. accuracy, AUROC, AP, R-squared, C-index).
Step 3: Evaluate the model
**************************
- **Evaluate on held-out set**: :ref:`Evaluate your final model <evaluation>` model on the held-out dataset.
Step 4: Generate heatmaps
*************************
- **Generate heatmaps**: :ref:`Generate heatmaps <generate_heatmaps>` of predictions across slides in the held-out dataset to assist with interpretability. For MIL models, heatmaps of both predictions and attention can be generated.
.. image:: heatmap_example.png
Step 5: Make a Mosaic map
*************************
- **Generate a mosaic map**: :ref:`Create a mosaic map <mosaic_map>`, which visually illustrates the latent space of your trained model and held-out dataset, to assist with interpretability.
.. image:: mosaic_example.png
Step 6: Live visualization
**************************
- **Deploy the model**: Finally, use a trained model to visualize predictions for whole-slide images with the interactive tool :ref:`Slideflow Studio <studio>`. This whole-slide image viewer includes deep learning tools enabling you to visualize model predictions on whole-slide images, standard JPG/PNG files, real-time camera feeds, and even Generative Adversarial Network (GAN)-generated images.
.. image:: workbench_preview.png