You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Before taking on the analysis of regions of interest we recommend to take a look at the tutorials for `watershed` and `threshold`.
24
24
:::
25
25
26
-
For the sake of simplicity we will use the same example as in a threshold method. However, we will put a little twist and use an image of [TIFF](https://en.wikipedia.org/wiki/TIFF'wikipedia link on .tiff format') format. This format is great for storing and editing images. It also allows adding metadata with extensive information about an image which we will examine a bit further in this tutorial. But for now let's take one step at a time.
26
+
For the sake of simplicity we will use the same example as in a threshold method. However, we will put a little twist and use an image of [TIFF](https://en.wikipedia.org/wiki/TIFF'wikipedia link on .tiff format') format. This format is great for storing and editing images of high quality. It also allows adding metadata with extensive information about an image which we will examine a bit further in this tutorial. But for now let's take one step at a time.
27
27
To get regions of interest you need to extract them from a map:
To do so we need to understand what kind of analysis is necessary. Depending on the answer different tools can be used. Let's say we want to find the filter regions by size and shape. Now,
51
-
For the size it is rather straight-forward. You can use the `getRois()` options, as was mentioned above, or you can use region's perimeter and surface properties to filter the ROIs.
48
+
To do so we need to understand what kind of analysis is necessary. Depending on the answer, different tools can be used. Let's say we want to filter regions by size and shape.
49
+
For filtering by size the process is rather straight-forward. You can use the `getRois()` options, as was mentioned above, or you can use region's perimeter and surface properties to filter the ROIs. We can detect
52
50
In this example let's get the regions which are above an average size of the `rois` sample.
53
-
First we need to find this average. It can be done in a rather straight-forward fashion:
51
+
First we need to find this average. It can be done like this:
The selected regions can be investigated further. For instance, we can use property like `roundness` to see how close the region's shape is to a circle. Let's put arbitrarily 0.9 as a limit (the coefficient for a perfect circle will be 1).
74
+
This provides us with a certain number of regions (colored in blue). The selected regions can be investigated further. For instance, we can use property like `roundness` to see how close the region's shape is to a circle. Let's put 0.9 as a limit (the coefficient for a perfect circle will be 1).
77
75
78
76
```ts
79
77
let roundestRois = [];
@@ -84,8 +82,6 @@ for (const roi of biggestRois) {
An image above highlights the ROIs that we found. Dark blue regions represent the particles that were above the average that we calculated. The light blue particles are the particles with an above average size and roundness above 0.9.
122
-
This is just a fraction of tools that ImageJS possesses. There are multiple properties that you can discover more about in our [API features](../Features/Regions%20of%20interest/Regions%20of%20interest.md) section. Here is an example of the properties that you can use with each region of interest.
This is just a fraction of tools that ImageJS possesses. There are many other properties that you can discover more about in our [API features](../Features/Regions%20of%20interest/Regions%20of%20interest.md) section. Here is an example of the properties that you can use with any region of interest:
Another aspect worth inspecting is extracting image metadata. If an image is of TIFF format, you can extract some metadata tags that can provide additional information about an image. For instance, you can get data such as image length and width or learn about image quality through bit depth(`bitsPerSample`) or X and Y Resolutions.
147
-
The metadata of TIFF format is split into two parts: `tiff` and `exif` which is another image format. It can provide information about camera settings, for example. `exif`, however, is mostly used for storing metadata within image formats like JPEG. And since an `exif` part of returned metadata is empty we will focus on `tiff` part.
141
+
Another aspect worth inspecting is extracting image metadata. Metadata represents information about an image itself. It can be something basic such as the date when an image was taken, or something specific like the name of the camera that the image was taken by. You can extract some metadata tags that can provide additional information about an image by using this command:
142
+
143
+
```ts
144
+
const meta =image.meta;
145
+
```
146
+
147
+
148
+
149
+
In ImageJS there are two supported formats for metadata: `exif` and `tiff`. While `exif` is a format used by digital images to store metadata, `tiff` is an image format used for high quality [raster graphics](https://en.wikipedia.org/wiki/Raster_graphics'wikipedia link for raster graphics') images. Since our image in question is of `tiff` format and `exif` part of returned metadata is empty we will focus on `tiff` part.
148
150
149
151
```ts
150
152
const meta =image.meta.tiff;
151
153
```
152
154
153
-
There you will have other two parts: one part will be comprised of a map with fields and then an object of TIFF meta tags which these fields' values are attributed to.
155
+
There you will have two other parts: one part will be comprised of a map with fields and then an object of TIFF meta tags which these fields' values are attributed to.
Within metadata, you might be wondering what this huge mix of letters and numbers represents:
160
162
161
163
162
164
163
-
These are custom fields added with additional information about an image. For instance, in this case you can get information about the microscope that was used, or the magnification level or the electrometric tension that was used while the image was taken. However, this data needs to be parsed, because it is difficult to decipher in its raw format.
164
-
To do so you need to identify what is the key id of this text. In our case it is 34682, but it might not always be the case so check it beforehand.
165
+
These are custom fields added with additional information about an image. For instance, in this case you can get information about the microscope that was used, or the magnification level or the electrometric tension that was used while the image was taken. However, this data needs to be parsed, as it is difficult to decipher in its raw format.
166
+
To do so you need to identify what is the key id of this text. In our case it is `34682`, but it might not always be the case so check it beforehand.
165
167
166
168
Next thing we need to do is to parse this text.
167
169
@@ -187,14 +189,14 @@ lines.forEach((a) => {
187
189
188
190
With this the data in the. console should look something like this.
189
191
190
-
192
+
191
193
192
194
### Getting pixel size
193
195
194
196
In this specific scenario we would also like to tell you about the way to calculate image's pixel size. It is an important aspect to deduce image's detail sharpness and display's quality.
195
-
Pixel size can be one of metadata fields but if this isn't the case we would like to show you how you can calculate it from existing data.
197
+
Pixel size can be one of metadata fields but if this isn't the case we would like to show you how you can calculate it from the existing data in this specific scenario.
196
198
197
-
If there is no such field as "Pixel size" you can calculate DPI resolution and apply it with magnification.
199
+
To calculate pixel size you can calculate DPI resolution and apply it with magnification.
198
200
DPI resolution represents the number of dots per inch. To calculate it we need to look at three lines in our parsed extra data: `XResolution`, `YResolution` and `ResolutionUnit`.
199
201
X and Y resolutions are the number of dots per inch on X and Y axes. So, if they are equal, then DPI resolution equals to one of these values. However, this value might not be measured in inches. To check that we need to look at the value of `ResolutionUnit`.
200
202
If its value equals to 2 then the X and Y resolutions are measured in inches.If it's 3 then it's in centimeters and has to be converted.
0 commit comments