Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

Kadir Cumali

PROJECT 6
Q.1.) In the project, object recognition has three steps, which are segmentation of target object,
finding the features of the segmented object, and matching the model features. In segmentation
part, it is implemented an algorithm to find the maximum contour area considering the threshold
value. It is obtained a segmented region that has a background. In order to get rid of the background
in large, it is implemented GrabCut algorithm. In this algorithm, it is requested to mark a single
rectangle around the object (that is max contour here), defining the outer part of the rectangle as
definite background, and the inner part of it as an unknown combination of the object (foreground)
and some background.

Figure 1: Segmented area(max contour) and segmented image


In the second part of the project, we need to find the features of the segmented object and
image to be compared by using SURF. We should detect the keypoints for the segmented object and
the image to be compared. After that, we should calculate descriptors (feature vectors). In matching
part, the descriptors are matched.

Figure 2: The result for the very similar images

Figure 3: Segmented area (threshold value:155)

Figure 4: The result for the slightly different images

Figure 5: Segmented area (threshold value:135)

Figure 6: The result for similar images, yet with substantial changes

Figure 7: The result for different images

You might also like