def __init__(self):
self.operate_dir = "../../saved_original_for_refraction/2/"
#self.operate_dir = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/original/cancer-CSCAN15-BSCAN15-spaing10um-02.jpg"
self.savedir_path = "../../saved_processed/2/"
#self.savedir_path = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/correct/hyper-bottom.jpg"
self.display_flag = True
from A_line import A_line_process
self.aline = A_line_process()
self.bias = 50
self.aligh_flag = True
self.rc_corrector = RC_function()
def __init__(self):
#self.operate_dir = "../../saved_original_for_refraction/image433.jpg"
self.operate_dir = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/original/Pedun-02-topview-cscan10-bscan15 -aver4 fr110.tif"
self.savedir_path = "../../saved_processed/"
self.savedir_path = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/correct/Pedun-02-topview-cscan10-bscan15 -aver4 fr110.tif"
self.display_flag = True
from A_line import A_line_process
self.aline = A_line_process()
self.bias = 50
self.aligh_flag = True
self.rc_corrector = RC_function()
class Seg_One_Frame(object):
def __init__(self):
self.operate_dir = "../../saved_original_for_refraction/2/"
#self.operate_dir = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/original/cancer-CSCAN15-BSCAN15-spaing10um-02.jpg"
self.savedir_path = "../../saved_processed/2/"
#self.savedir_path = "D:/PhD/IPB meeting/natalia paper/fractional correction/Bscan/correct/hyper-bottom.jpg"
self.display_flag = True
from A_line import A_line_process
self.aline = A_line_process()
self.bias = 50
self.aligh_flag = True
self.rc_corrector = RC_function()
#input image to calculate the fist aveline, for get the local minimal value calculation
def calculate_the_average_line(self, img):
calculate_w = 5 # only calculate the front10 lines
H, W = img.shape
new = img[:, 5:5 + calculate_w]
ave_line = new.sum(axis=1) / calculate_w
return ave_line
#the validation functionfor check the matrix and can also be used for validate the correction result
def seg_process(self, Img_ini):
gray = Img_ini
H, W = gray.shape
#gray = gray [: ,100: W -100]
#Img = Img [: ,100: W -100]
ave_line = self.calculate_the_average_line(gray)
peaks = self.aline.find_4peak(ave_line)
H, W = gray.shape
original = gray
#gray = cv2.blur(gray,(3,3))
#gray = cv2.medianBlur(gray,5)
#gray = cv2.blur(gray,(5,5))
gray = cv2.GaussianBlur(gray, (3, 3), 0)
#gray = cv2.bilateralFilter(gray,15,75,75)
#gray = cv2.bilateralFilter(gray,15,75,75)
ave_line = self.calculate_the_average_line(gray)
peaks = self.aline.find_4peak(ave_line)
#gray = cv2.blur(gray,(5,5),0)
if self.display_flag == True:
cv2.imshow('ini', Img_ini)
cv2.imshow('blur', gray.astype(np.uint8))
x_kernel = np.asarray([
[-1, 0, 1], # Sobel kernel for x-direction
[-2, 0, 2],
[-1, 0, 1]
])
#y_kernel = np.asarray([[-2, -2, -2], # Sobel kernel for y-direction
# [1, 1, 1],
# [1, 1, 1]])
y_kernel = np.asarray([
[-1, -1, -1], # Sobel kernel for y-direction
[-1, -1, -1],
[-1, -1, -1],
[6, 6, 6],
[-1, -1, -1],
[-1, -1, -1],
[-1, -1, -1]
])
#y_kernel = np.asarray([ # Sobel kernel for y-direction
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [12,12],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# ])
y_kernel = np.asarray([ # Sobel kernel for y-direction
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[28, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
])
y_kernel = y_kernel / 9
gray = gray.astype(np.float)
#sobel_x = signal.convolve2d(gray, x_kernel) #
sobel_y = signal.convolve2d(gray,
y_kernel) # convolve kernels over images
sobel_y = np.clip(sobel_y + 10, 1, 254)
sobel_y = cv2.medianBlur(sobel_y.astype(np.uint8), 5)
sobel_y = cv2.GaussianBlur(sobel_y, (5, 5), 0)
sobel_y = cv2.blur(sobel_y, (5, 5))
#sobel_y = cv2.GaussianBlur(sobel_y,(5,5),0)
#sobel_y = cv2.bilateralFilter(sobel_y.astype(np.uint8),9,175,175)
#find the start 4 point
#sobel_y=sobel_y[self.bias:H-self.bias, :]
#Img=Img[self.bias:H-self.bias, :]
ave_line = self.calculate_the_average_line(sobel_y)
peaks = self.aline.find_4peak(ave_line)
Rever_img = 255 - sobel_y
#start_point = 596
peaks = np.clip(peaks, 1, Rever_img.shape[0] - 1)
#new_peak = np.zeros(4)
#new_peak=peaks
#new_peak[3] = peaks[2]+35
#peaks = new_peak
if Manual_start_flag == True:
peaks[1] = peaks[0] + 472 - 293
peaks[2] = peaks[0] + 500 - 293
peaks[3] = peaks[0] + 677 - 293
path1, path_cost1 = PATH.search_a_path(Rever_img, int(peaks[0]))
#path2,path_cost1=PATH.search_a_path(Rever_img,int(peaks[1]))
#path3,path_cost1=PATH.search_a_path(Rever_img,int(peaks[2]))
#path4,path_cost1=PATH.search_a_path(Rever_img,int(peaks[3]))
path1 = gaussian_filter1d(path1, 6)
#path2 =gaussian_filter1d(path2,2)
#path3 =gaussian_filter1d(path3,2)
#path4 =gaussian_filter1d(path4,2)
#path4=path3
##path2 = path3
#path3 = path2+35
#path2 = path2 - 5
#path3,path_cost1=PATH.search_a_path_based_on_path(Rever_img,path3)
#path3 =gaussian_filter1d(path3,4)
path1 = path1 #-10
path1 = np.clip(path1, 0, sobel_y.shape[0] - 1)
#[path1,path2,path3,path4]=np.clip([path1,path2,path3,path4],
# 0,sobel_y.shape[0]-1)
for i in range(len(path1)):
sobel_y[int(path1[i]), i] = 254
Dark_boundaries = sobel_y * 0
path1 = np.clip(path1, 0, Dark_boundaries.shape[0] - 2)
for i in range(len(path1)):
Dark_boundaries[int(path1[i]), i] = 254
for i in range(gray.shape[1]):
gray[int(path1[i]) + 1, i] = gray[int(path1[i]), i] = 254
# corect tje RC
#original = cv2.resize(original, (int(W/2),int(H/2)), interpolation=cv2.INTER_LINEAR)
new = self.rc_corrector.correct(original, path1)
if self.display_flag == True:
cv2.imshow('revert', Rever_img.astype(np.uint8))
#cv2.imshow('path on blur',gray.astype(np.uint8))
cv2.imshow('Seg2', Img)
#sobel_y = sobel_y*0.1
#edges = cv2.Canny(gray,50, 300,10)
cv2.imshow('seg', sobel_y.astype(np.uint8))
cv2.imshow('correct', new.astype(np.uint8))
#cv2.imwrite(self.savedir_path + str(1) +".jpg",sobel_y .astype(np.uint8))
cv2.waitKey(1)
return new # Img,sobel_y,Dark_boundaries,[path1,path2,path3,path4]
def reference(self, Img):
gray = Img
H, W = gray.shape
#gray = gray [: ,100: W -100]
#Img = Img [: ,100: W -100]
ave_line = self.calculate_the_average_line(gray)
peaks = self.aline.find_4peak(ave_line)
H, W = gray.shape
original = gray
#gray = cv2.blur(gray,(3,3))
#gray = cv2.medianBlur(gray,5)
#gray = cv2.blur(gray,(5,5))
gray = cv2.GaussianBlur(gray, (3, 3), 0)
#gray = cv2.bilateralFilter(gray,15,75,75)
#gray = cv2.bilateralFilter(gray,15,75,75)
ave_line = self.calculate_the_average_line(gray)
peaks = self.aline.find_4peak(ave_line)
#gray = cv2.blur(gray,(5,5),0)
if self.display_flag == True:
cv2.imshow('ini', Img)
cv2.imshow('blur', gray.astype(np.uint8))
x_kernel = np.asarray([
[-1, 0, 1], # Sobel kernel for x-direction
[-2, 0, 2],
[-1, 0, 1]
])
#y_kernel = np.asarray([[-2, -2, -2], # Sobel kernel for y-direction
# [1, 1, 1],
# [1, 1, 1]])
y_kernel = np.asarray([
[-1, -1, -1], # Sobel kernel for y-direction
[-1, -1, -1],
[-1, -1, -1],
[6, 6, 6],
[-1, -1, -1],
[-1, -1, -1],
[-1, -1, -1]
])
#y_kernel = np.asarray([ # Sobel kernel for y-direction
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [12,12],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# [-1,-1],
# ])
y_kernel = np.asarray([ # Sobel kernel for y-direction
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[28, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
[-1, 0],
])
y_kernel = y_kernel / 9
gray = gray.astype(np.float)
#sobel_x = signal.convolve2d(gray, x_kernel) #
sobel_y = signal.convolve2d(gray,
y_kernel) # convolve kernels over images
sobel_y = np.clip(sobel_y + 10, 1, 254)
sobel_y = cv2.medianBlur(sobel_y.astype(np.uint8), 5)
sobel_y = cv2.GaussianBlur(sobel_y, (5, 5), 0)
sobel_y = cv2.blur(sobel_y, (5, 5))
#sobel_y = cv2.GaussianBlur(sobel_y,(5,5),0)
#sobel_y = cv2.bilateralFilter(sobel_y.astype(np.uint8),9,175,175)
#find the start 4 point
sobel_y = sobel_y[self.bias:H - self.bias, :]
Img = Img[self.bias:H - self.bias, :]
ave_line = self.calculate_the_average_line(sobel_y)
peaks = self.aline.find_4peak(ave_line)
Rever_img = 255 - sobel_y
#start_point = 596
peaks = np.clip(peaks, 1, Rever_img.shape[0] - 1)
#new_peak = np.zeros(4)
#new_peak=peaks
#new_peak[3] = peaks[2]+35
#peaks = new_peak
if Manual_start_flag == True:
peaks[1] = peaks[0] + 472 - 293
peaks[2] = peaks[0] + 500 - 293
peaks[3] = peaks[0] + 677 - 293
path1, path_cost1 = PATH.search_a_path(Rever_img, int(peaks[0]))
#path2,path_cost1=PATH.search_a_path(Rever_img,int(peaks[1]))
#path3,path_cost1=PATH.search_a_path(Rever_img,int(peaks[2]))
#path4,path_cost1=PATH.search_a_path(Rever_img,int(peaks[3]))
path1 = gaussian_filter1d(path1, 2)
#path2 =gaussian_filter1d(path2,2)
#path3 =gaussian_filter1d(path3,2)
#path4 =gaussian_filter1d(path4,2)
#path4=path3
##path2 = path3
#path3 = path2+35
#path2 = path2 - 5
#path3,path_cost1=PATH.search_a_path_based_on_path(Rever_img,path3)
#path3 =gaussian_filter1d(path3,4)
path1 = path1 #-10
path1 = np.clip(path1, 0, sobel_y.shape[0] - 1)
#[path1,path2,path3,path4]=np.clip([path1,path2,path3,path4],
# 0,sobel_y.shape[0]-1)
for i in range(len(path1)):
sobel_y[int(path1[i]), i] = 254
Dark_boundaries = sobel_y * 0
path1 = np.clip(path1, 0, Dark_boundaries.shape[0] - 2)
for i in range(len(path1)):
Dark_boundaries[int(path1[i]), i] = 254
for i in range(Img.shape[1]):
Img[int(path1[i]) + 1, i] = Img[int(path1[i]), i] = 254
# corect tje RC
#original = cv2.resize(original, (int(W/1),int(H/1)), interpolation=cv2.INTER_LINEAR)
#new = self.rc_corrector.correct(original,path1)
if self.display_flag == True:
cv2.imshow('revert', Rever_img.astype(np.uint8))
#cv2.imshow('path on blur',gray.astype(np.uint8))
cv2.imshow('Seg2', Img)
#sobel_y = sobel_y*0.1
#edges = cv2.Canny(gray,50, 300,10)
cv2.imshow('seg', sobel_y.astype(np.uint8))
#cv2.imshow('correct',new.astype(np.uint8))
#cv2.imwrite(self.savedir_path + str(1) +".jpg",sobel_y .astype(np.uint8))
cv2.waitKey(1)
return path1