Page 214 - Ai_C10_Flipbook
P. 214
After completing the activity, we conclude that,
In image processing, a wide variety of features can be extracted from an image, such as blobs, edges, and corners.
These features play a crucial role in enabling various tasks, such as object detection, image segmentation, and
pattern recognition. The analysis derived from these features depends on the specific application at hand.
A fundamental question that often arises is: which features are most effective for a given task? As demonstrated
in the previous activity, corners are particularly valuable features because they are unique and can be identified at
specific locations in an image. In contrast, edges, while also useful, are spread across a line and tend to look the
same along their length, making them less distinctive for precise localisation.
This information highlights that corners are excellent features to extract from
an image, as they provide precise and reliable information. Edges, while
slightly less distinctive, are still beneficial and can be used to complement
corner-based analysis. Additionally, combining these features with other
attributes, such as texture and intensity variations, can further enhance the
accuracy and effectiveness of image processing tasks.
Let us understand the concept of image feature with the help of another
activity.
Observe the images provided and apply the idea of identifying good features in an image.
In the image above, how can we accurately determine the exact location of every patch? Here’s how different
patches behave:
• Blue Patch: This is a flat area, making it challenging to locate and track. No matter where you move the blue
patch within the image, it appears the same, offering no distinct features for identification.
• Black Patch: This represents an edge in the image. If you move the black patch along the edge (parallel to it),
the appearance remains unchanged, making it difficult to determine its precise position.
• Red Patch: This corresponds to a corner in the image. Unlike the other patches, wherever you move the red
patch, it looks different, making it unique and easily identifiable.
Corners, like the red patch, are considered good features in an image because they are distinct and can be
accurately tracked or located.
Reboot
1. In image processing, which type of feature is most commonly used for tracking objects in an image?
2. What could be the benefit of combining features like texture and intensity with edges and corners in
image processing?
Convolution
Till now, we have learned an image is a visual representation, typically made up of tiny elements called pixels.
When many pixels are arranged together in a grid, they form a complete image that we can see and recognise.
Each pixel has a value ranging from 0 to 255 to specify color and brightness.
The computer stores and image in number. But what happens if we modify these numbers? The answer is simple:
the image changes. Altering pixel values directly impacts how the image looks, and this concept forms the
foundation of image editing.
212 Artificial Intelligence Play (Ver 1.0)-X

