Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd Logo
Upload

Welcome to Scribd!

  • 28 views

    Dip Lab - 9

    Uploaded by

    Golam Daiyan
    This document describes a lab assignment on image segmentation using k-means clustering. The objective is to implement cluster-based segmentation of an image. Students are expected to complete two exercises: 1) load an image and perform color-based segmentation into 5 clusters using k-means clustering, and 2) load another image and perform convolution with 4 line detection masks to identify lines. The tasks are to be demonstrated to the instructor for signing off.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Dip Lab - 9

    Uploaded by

    Golam Daiyan
    28 views3 pages
    This document describes a lab assignment on image segmentation using k-means clustering. The objective is to implement cluster-based segmentation of an image. Students are expected to complete two exercises: 1) load an image and perform color-based segmentation into 5 clusters using k-means clustering, and 2) load another image and perform convolution with 4 line detection masks to identify lines. The tasks are to be demonstrated to the instructor for signing off.

    Original Description:

    Lab3

    Original Title

    Dip Lab -9

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document describes a lab assignment on image segmentation using k-means clustering. The objective is to implement cluster-based segmentation of an image. Students are expected to complete two exercises: 1) load an image and perform color-based segmentation into 5 clusters using k-means clustering, and 2) load another image and perform convolution with 4 line detection masks to identify lines. The tasks are to be demonstrated to the instructor for signing off.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    28 views3 pages

    Dip Lab - 9

    Uploaded by

    Golam Daiyan
    This document describes a lab assignment on image segmentation using k-means clustering. The objective is to implement cluster-based segmentation of an image. Students are expected to complete two exercises: 1) load an image and perform color-based segmentation into 5 clusters using k-means clustering, and 2) load another image and perform convolution with 4 line detection masks to identify lines. The tasks are to be demonstrated to the instructor for signing off.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 3

    CEN 440 Digital Image Processing

    Implementation of cluster based segmentation of an image


    Lab 9
    Name: _________________________________

    Enrollment #: _________________________________

    Class: _________________________________

    Objective

    Today’s lab covers segmentation of images into regions. We will be specifically focusing on
    segmentation based on clustering using the classical k-means clustering algorithm. By the end of
    this lab you should be able to implement cluster based segmentation of an image.

    Submission Requirements

    You are expected to complete the assigned tasks within the lab session and show them to the
    lab engineer/instructor. Get the lab journal signed by your instructor and submit it by the end of
    lab session.

    Tasks

    Segmentation divides an image into disjoint regions such that pixels within a region share some
    common attributes. Segmentation algorithms are broadly categorized into similarity and
    discontinuity based methods. Similarity based segmentation exploits the key idea that pixels
    within a region are ‘similar’ to one another. Discontinuity based methods, on the other hand,
    rely on finding object boundaries and edges to identify different regions in an image.

    A common approach in similarity based segmentation is to cluster similar pixels into groups
    using standard clustering algorithms. The most commonly used of these clustering algorithms is
    the k-mean clustering where given the number of clusters k, the algorithm iteratively partitions
    the given data into k groups or clusters. In case of images, each pixel can be represented as a
    point in a three dimensional space. The clustering algorithm can then be applied to identify

    Digital Image Processing Page 1


    Lab Journal – Lab 9
    clusters of pixels in the color space and producing k distinct regions. Figure 1 illustrates an
    example of applying k-mean algorithm to an image.

    Figure 1 An image segmented into three clusters

    The statistics tool box in Matlab provides the kmeans function to cluster data into k groups. The
    function requires the data matrix and number of clusters as input and returns the index (cluster
    number) of each data member. In case of colored images, we can extract the individual
    channels of the image and create the data matrix as follows.

    dataMatrix=[R;G;B];
    dataMatrix =double( dataMatrix');

    Where R, G and B represent the red, green and blue channels of the image each represented as
    a vector. The data matrix fed to the k-means functions requires each row to be one data point.
    Hence, the size of data matrix is M x 3 where M is the total number of pixels in the image. The
    function returns an Mx1 vector giving the cluster number to which each pixel belongs. Each
    cluster can then be represented by a different color to represent a region/cluster.

    Digital Image Processing Page 2


    Lab Journal – Lab 9

    Exercise 1

    Load the image ‘onion.png’ and perform color based segmentation using k-means
    clustering. Provide k = 5 and represent each cluster by a different color.

    In addition to similarity based segmentaiton, discontinuities in the image can also be exploited
    to segment objects. Convolution with filters based on image derivates can be used for this
    purpose. These filters can be employed to detect points, edges and lines in an image. In the
    following exercise, you will detect lines in a givn image using line detection masks.

    Exercise 2

    Load the provided image ‘field.png’ and perform convolution with the four line detection
    masks. Display the four output images to see the response of each filter on the image.

    Do Exercises 1 and 2 and get them checked by your instructor. If you are unable to complete the
    tasks in the lab session, deposit this sheet alongwith your programs (printed or handwritten)
    before the start of the next lab session.

    S No. Exercise Checked By:


    1. Exercise 1
    2. Exercise 2

    +++++++++++++++++++++++++++

    Digital Image Processing Page 3

    Powered by TCPDF (www.tcpdf.org)

    You might also like