Skip to content

Latest commit

 

History

History
123 lines (99 loc) · 3.27 KB

Task4.md

File metadata and controls

123 lines (99 loc) · 3.27 KB

Task 4

Thresholding

Global Atsu
std::vector<int> globalAtsu(Image &inputImg, int histSize = 256, int numModes = 2);

function parameters:

  • inputImg: reference to gray Image
  • histSize: histogram size depending on the number of bits that represent each pixel in the image
  • numModes: number of modes

function return: return array of thresholds depending on number of modes

Results:

  • global atsu 2 modes

  • global atsu 3 modes

Local Atsu
Image localAtsu(Image &inputImg, int blockDim, int histSize = 256, int numModes = 2);

function parameters:

  • inputImg: reference to gray Image
  • blockDim: block dimension (eg, 5 means 5 x 5)
  • histSize: histogram size depending on the number of bits that represent each pixel in the image
  • numModes: number of modes

function return: return the thresholded image

Results:

  • local atsu 2 modes, blockDim: 5

  • local atsu 3 modes, blockDim: 5

Global Optimal Iterative Thresholding
int globalOptimalIterativeThresholding(Image &inputImg);

function parameters:

  • inputImg: reference to gray Image

function return: return threshold

Results:

Local Optimal Iterative Thresholding
Image localOptimalIterativeThresholding(Image &inputImg, int blockDim);

function parameters:

  • inputImg: reference to gray Image
  • blockDim: block dimension (eg, 5 means 5 x 5)

function return: return threshold

Results:
blockDim: 7

Means-Shift
MeanShift(Image *image, float hs, float hr)

class functions:

  • c++ float distance(int i1, int j1, int i2, int j2
  • c++ float distance(vector<float> pt1, vector<float> pt2)
  • c++ Image run()

Results:
blockDim: 7

K-Mean
K_mean(Image *image, int k, int maxIteration)

class functions:

  • c++ void clusterPixels()
  • c++ int closestCluster(int i, int j)
  • c++ void getClustersCentroid()
  • c++ bool centroidsChanged(vector<vector<float>> oldCentroids)
  • c++ float distance(int i, int j, vector<float> centroid)
  • c++ void run()
  • c++ K_mean()
  • c++ Image getOutput()
Region Growing
Image RGSegmentation(Image & inputImg, std::vector<point> seedPoint)

--> You should select seed points from image in gui to segment the regions of these seed points

function parameters:

  • inputImg: reference to color image
  • seedPoint: Points which are used as seeds for segmentation regions

function return: return segmented output image Results:
blockDim: 7

Agglomerative
Image agglomerativeSeg(Image & img, int numOfClusters, unsigned long initialClustersNum)

function parameters:

  • img: reference to gray image
  • numOfClusters : the number of clusters needed in the output image
  • initialClustersNum : The initial numbers of clusters which is merged until reach to the the numOfClusters

function return: return threshold Results:
blockDim: 7