Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Connected components suggestions #316

Merged
merged 6 commits into from
Dec 15, 2023
Merged

Connected components suggestions #316

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
b32977f
display range text
JeremyPike Dec 12, 2023
b2a5044
typo
JeremyPike Dec 12, 2023
bbcc281
naming function inputs
JeremyPike Dec 12, 2023
50f56db
refactoring solution
JeremyPike Dec 13, 2023
57fbf84
rewording
JeremyPike Dec 13, 2023
f072c71
Updated spoiler heading
JeremyPike Dec 15, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
53 changes: 30 additions & 23 deletions episodes/08-connected-components.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -316,24 +316,22 @@ labeled_image, count = connected_components(filename="data/shapes-01.jpg", sigma

fig, ax = plt.subplots()
plt.imshow(labeled_image)
plt.axis("off");
plt.axis("off")
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I would keep the semicolon, so the output is not cluttered with text and memory addresses... But maybe the semicolon should be explained to learners? cf. mkcor/python-prog#8 (comment) 😉

Copy link
Contributor Author

@JeremyPike JeremyPike Dec 18, 2023 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sure, sorry, I probably shouldn't have removed this! I don't use Jupyter much apart from for teaching so from my perspective I thought it might be confusing for some participants and I didn't mind seeing the limits of the axis printed. Happy to put back in with a brief explanation :)

```

:::::::::::::::: spoiler

## Color mappings
## Do you see an empty image?

Here you might get a warning
If you are using an old version of Matplotlib you might get a warning
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
If you are using an old version of Matplotlib you might get a warning
If you are using an older version of Matplotlib you might get a warning

`UserWarning: Low image data range; displaying image with stretched contrast.`
or just see an all black image
(Note: this behavior might change in future versions or
not occur with a different image viewer).
or just see a visually empty image.

What went wrong?
When you hover over the black image,
When you hover over the image,
the pixel values are shown as numbers in the lower corner of the viewer.
You can see that some pixels have values different from `0`,
so they are not actually pure black.
so they are not actually all the same value.
Let's find out more by examining `labeled_image`.
Properties that might be interesting in this context are `dtype`,
the minimum and maximum value.
Expand All @@ -345,29 +343,36 @@ print("min:", np.min(labeled_image))
print("max:", np.max(labeled_image))
```

Examining the output can give us a clue why the image appears black.
Examining the output can give us a clue why the image appears empty.

```output
dtype: int32
min: 0
max: 11
```

The `dtype` of `labeled_image` is `int64`.
This means that values in this image range from `-2 ** 63` to `2 ** 63 - 1`.
The `dtype` of `labeled_image` is `int32`.
This means that values in this image range from `-2 ** 31` to `2 ** 31 - 1`.
Those are really big numbers.
From this available space we only use the range from `0` to `11`.
When showing this image in the viewer,
it squeezes the complete range into 256 gray values.
Therefore, the range of our numbers does not produce any visible change.
it may squeeze the complete range into 256 gray values.
Therefore, the range of our numbers does not produce any visible variation. One way to rectify this
is to explicitly specify the data range we want the colormap to cover:

Fortunately, the scikit-image library has tools to cope with this situation.
```python
fig, ax = plt.subplots()
plt.imshow(labeled_image, vmin=np.min(labeled_image), vmax=np.max(labeled_image))
Comment on lines +364 to +365
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
fig, ax = plt.subplots()
plt.imshow(labeled_image, vmin=np.min(labeled_image), vmax=np.max(labeled_image))
plt.imshow(labeled_image, vmin=np.min(labeled_image), vmax=np.max(labeled_image))

Since fig and ax are not being used?

Copy link
Contributor Author

@JeremyPike JeremyPike Dec 18, 2023 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fig, ax = plt.subplots() is used here as I wanted to create a new figure rather than use the previous one, its used throughout the course even when fig and ax are not used so I'm hesitant to change to an alternative or remove the line. Maybe I don't need to create a new figure though, in which case the line can be removed as suggested?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Oh, ok, thanks for explaining! So, I would suggest you keep this line but replace the next one with:

ax.imshow(labeled_image, vmin=np.min(labeled_image), vmax=np.max(labeled_image))

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree this is better but its not consistent with how its done in the rest of the course so perhaps this should be a different PR where all similar examples are changed at once.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I agree; I believe that we should use this "object-oriented (OO) style" throughout the lesson, since we create figure and axis objects now. At the moment, we are mixing up the OO-style with the pyplot-style [1]. For instance, 328d544 adds figure and axis objects, but we still have all the pyplot-style legacy. I would change, e.g.,

fig, ax = plt.subplots()
plt.plot(bin_edges[0:-1], histogram)
plt.title("Grayscale Histogram")
plt.xlabel("grayscale value")
plt.ylabel("pixels")
plt.xlim(0, 1.0)

into

fig, ax = plt.subplots()
ax.plot(bin_edges[0:-1], histogram)
ax.set_title("Grayscale Histogram")
ax.set_xlabel("grayscale value")
ax.set_ylabel("pixels")
ax.set_xlim(0, 1.0);

/cc @datacarpentry/image-processing-curriculum-maintainers

[1] https://matplotlib.org/stable/users/explain/quick_start.html#the-explicit-and-the-implicit-interfaces

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch, @mkcor. Please could you open a new issue where we can track this?

```

Note this is the default behaviour for newer versions of `matplotlib.pyplot.imshow`.
Alternatively we could convert the image to RGB and then display it.


:::::::::::::::::::::::::

We can use the function `ski.color.label2rgb()`
to convert the colours in the image
to convert the 32-bit grayscale labeled image to standard RGB colour
(recall that we already used the `ski.color.rgb2gray()` function
to convert to grayscale).
With `ski.color.label2rgb()`,
Expand All @@ -380,7 +385,7 @@ colored_label_image = ski.color.label2rgb(labeled_image, bg_label=0)

fig, ax = plt.subplots()
plt.imshow(colored_label_image)
plt.axis("off");
plt.axis("off")
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Same comment as above.

```

![](fig/shapes-01-labeled.png){alt='Labeled objects'}
Expand All @@ -402,7 +407,7 @@ How does changing the `sigma` and `threshold` values influence the result?
## Solution

As you might have guessed, the return value `count` already
contains the number of found images. So it can simply be printed
contains the number of found objects in the image. So it can simply be printed
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
contains the number of found objects in the image. So it can simply be printed
contains the number of objects found in the image. So it can simply be printed

with

```python
Expand Down Expand Up @@ -685,7 +690,7 @@ set the entries that belong to small objects to `0`.
object_areas = np.array([objf["area"] for objf in object_features])
object_labels = np.array([objf["label"] for objf in object_features])
small_objects = object_labels[object_areas < min_area]
labeled_image[np.isin(labeled_image,small_objects)] = 0
labeled_image[np.isin(labeled_image, small_objects)] = 0
```

An even more elegant way to remove small objects from the image is
Expand All @@ -698,7 +703,7 @@ i.e, their pixel values are set to `False`.
We can then apply `ski.measure.label` to the masked image:

```python
object_mask = ski.morphology.remove_small_objects(binary_mask,min_area)
object_mask = ski.morphology.remove_small_objects(binary_mask, min_size=min_area)
labeled_image, n = ski.measure.label(object_mask,
connectivity=connectivity, return_num=True)
```
Expand All @@ -712,7 +717,7 @@ def enhanced_connected_components(filename, sigma=1.0, t=0.5, connectivity=2, mi
gray_image = ski.color.rgb2gray(image)
blurred_image = ski.filters.gaussian(gray_image, sigma=sigma)
binary_mask = blurred_image < t
object_mask = ski.morphology.remove_small_objects(binary_mask,min_area)
object_mask = ski.morphology.remove_small_objects(binary_mask, min_size=min_area)
labeled_image, count = ski.measure.label(object_mask,
connectivity=connectivity, return_num=True)
return labeled_image, count
Expand All @@ -728,7 +733,7 @@ colored_label_image = ski.color.label2rgb(labeled_image, bg_label=0)

fig, ax = plt.subplots()
plt.imshow(colored_label_image)
plt.axis("off");
plt.axis("off")

print("Found", count, "objects in the image.")
```
Expand Down Expand Up @@ -772,14 +777,16 @@ the area by indexing the `object_areas` with the label values in `labeled_image`

```python
object_areas = np.array([objf["area"] for objf in ski.measure.regionprops(labeled_image)])
object_areas = np.insert(0,1,object_areas)
# prepend zero to object_areas array for background pixels
object_areas = np.insert(0, obj=1, values=object_areas)
# create image where the pixels in each object are equal to that object's area
colored_area_image = object_areas[labeled_image]

fig, ax = plt.subplots()
im = plt.imshow(colored_area_image)
cbar = fig.colorbar(im, ax=ax, shrink=0.85)
cbar.ax.set_title("Area")
plt.axis("off");
plt.axis("off")
```

![](fig/shapes-01-objects-coloured-by-area.png){alt='Objects colored by area'}
Expand Down