Segmentation #4
Merged
Segmentation #4
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mkcor
reviewed
Aug 26, 2024
|
|
||
| Think of Michaelangelo, carving the statue of David. The block of marble is *segmented* into foreground and background. In this case, the background is discarded, and only the work of art remains. | ||
|
|
||
| To be useful, segmentation has to be followed by *measurement* and *interpretation*. We measure attributes of the segments, and then interpret those measurements to ask some question. Ultimately, we want to know: are the cells healthy, or is the part manufactured correctly, or is the land flooded. |
There was a problem hiding this comment.
Suggested change
| To be useful, segmentation has to be followed by *measurement* and *interpretation*. We measure attributes of the segments, and then interpret those measurements to ask some question. Ultimately, we want to know: are the cells healthy, or is the part manufactured correctly, or is the land flooded. | |
| To be useful, segmentation has to be followed by *measurement* and *interpretation*. We measure attributes of the segments, and then interpret those measurements to answer some question. Ultimately, we want to know: are the cells healthy, or is the part manufactured correctly, or is the land flooded. |
or 'try and answer'
|
|
||
| Let's consider that second block, the segmentation algorithm, for a moment. | ||
|
|
||
| Scikit-image implements several *heuristic* algorithms for segmentation. I.e., these are typically *unsupervised*, based on rules around properties of the image / pixels, rather than on labeling examples of objects and non-objects. Neural network approaches, such as U-Net, DeepLab, and Mask R-CNN have proved very effective for segmenting images, given enough training data. These are likely what would be used in practice, unless sufficient labeled data is unavailable. |
There was a problem hiding this comment.
The double negative is twisting my brains... "unless not enough labeled data are available?"
|
|
||
| How shall we go about representing segmentations? We already know how to use `numpy` to represent images. Can we use the same concept here? | ||
|
|
||
| Yes, we can create a *label* image, of the same size as the segmented image, with each value getting a different value. |
There was a problem hiding this comment.
Suggested change
| Yes, we can create a *label* image, of the same size as the segmented image, with each value getting a different value. | |
| Yes, we can create a *label* image, of the same shape as the segmented image, with each segment getting a different value. |
Maybe a "labels" image/array? E.g., see
from typing import Annotated
ImageData = Annotated[np.ndarray, 'image']
LabelsData = Annotated[np.ndarray, 'labels']
CoordsData = Annotated[np.ndarray, 'coordinates']from https://github.com/scikit-image/skimage-archive/blob/main/grants/2022_CZI_EOSS5/full_application.md
| ax[2].set_title('labels'); | ||
| ``` | ||
|
|
||
| Notice that "labels" is just a NumPy array with integer values. We have to be careful to interpret it as labels and not as an image. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.