median_noise.py

import numpy as np
import pydicom
from algorithms.sorting import quick_sort
import pectoral_muscle

def noise_removal(DDSM):
    for i in (DDSM):
        ds = pydicom.dcmread(i)
        # initialize the y part of the pixel in the array
        y = 0
        # initialize x
        x = 0
        # size of image
        pixels = ds.pixel_array.shape[0] * ds.pixel_array.shape[1]
        for pixel in range(pixels):
            # define the pixel we're looking at
            """
            n = ds.pixel_array[y , x]
            n_up = ds.pixel_array[y - 1, x]
            n_upl = ds.pixel_array[y - 1, x - 1]
            n_l = ds.pixel_array[y, x - 1]
            n_downl = ds.pixel_array[y + 1, x - 1]
            n_down = ds.pixel_array[y + 1, x]
            n_downr = ds.pixel_array[y + 1, x + 1]
            n_r = ds.pixel_array[y, x + 1]
            n_upr = ds.pixel_array[y - 1, x + 1]

            window = [n, n_up, n_upr, n_upl, n_l, n_downl, n_down, n_downr, n_r]
            """
            window = pectoral_muscle.neighbors(y, x, ds.pixel_array)
                # sorting
            window = quick_sort.sort(window)
            # set value to pixel\
            ds.pixel_array[y, x] = ds.pixel_array[window[len(window) / 2][0], window[len(window) / 2][1]]
            if x == ds.pixel_array.shape[1] - 1:
                y = y + 1
                x = 0
                continue
            x += 1
            # print("done with iteration " + str(pixel) +" for median noise")
        ds.PixelData = ds.pixel_array.tostring()
        ds.save_as(i)

def noise_removal_single(i):
    ds = pydicom.dcmread(i)
    #initialize the y part of the pixel in the array
    y = 0
    #initialize x
    x = 0
    #size of image
    pixels = ds.pixel_array.shape[0]*ds.pixel_array.shape[1]
    for pixel in range(pixels):
        # define the pixel we're looking at
        """
        n = ds.pixel_array[y , x]
        n_up = ds.pixel_array[y - 1, x]
        n_upl = ds.pixel_array[y - 1, x - 1]
        n_l = ds.pixel_array[y, x - 1]
        n_downl = ds.pixel_array[y + 1, x - 1]
        n_down = ds.pixel_array[y + 1, x]
        n_downr = ds.pixel_array[y + 1, x + 1]
        n_r = ds.pixel_array[y, x + 1]
        n_upr = ds.pixel_array[y - 1, x + 1]
        
        window = [n, n_up, n_upr, n_upl, n_l, n_downl, n_down, n_downr, n_r]
        """
        window = pectoral_muscle.neighbors(y, x, ds.pixel_array)
        #sorting
        window = quick_sort.sort(window)
        #set value to pixel\
        ds.pixel_array[y, x] = ds.pixel_array[window[len(window)/2][0], window[len(window)/2][1]]
        if x == ds.pixel_array.shape[1]-1:
            y = y + 1
            x = 0
            continue
        x += 1
        #print("done with iteration " + str(pixel) +" for median noise")
    ds.PixelData = ds.pixel_array.tostring()
    ds.save_as(i)