Python实现PS滤镜Fish lens图像扭曲效果示例

import numpy as npfrom skimage import img_as_floatimport matplotlib.pyplot as pltfrom skimage import ioimport mathimport numpy.matlibfile_name2='D:/Visual Effects/PS Algorithm/4.jpg'img=io.imread(file_name2)img = img_as_float(img)row, col, channel = img.shapeimg_out = img * 1.0R=(min(row, col)/2)# gamma > 1 zoom in gamma < 1 zoom outgamma = 1.5center_x = (col-1)/2.0center_y = (row-1)/2.0xx = np.arange (col) yy = np.arange (row)x_mask = numpy.matlib.repmat (xx, row, 1)y_mask = numpy.matlib.repmat (yy, col, 1)y_mask = np.transpose(y_mask)xx_dif = x_mask - center_xyy_dif = center_y - y_maskr = np.sqrt(xx_dif * xx_dif + yy_dif * yy_dif)theta = np.arctan(yy_dif / xx_dif)mask_1 = xx_dif < 0theta = theta * (1 - mask_1) + (theta + math.pi) * mask_1r_new = R*np.power(r/R, gamma)x_new = r_new * np.cos(theta) + center_xy_new = center_y - r_new * np.sin(theta) int_x = np.floor (x_new)int_x = int_x.astype(int)int_y = np.floor (y_new)int_y = int_y.astype(int)for ii in range(row):  for jj in range (col):    new_xx = int_x [ii, jj]    new_yy = int_y [ii, jj]    if x_new [ii, jj] < 0 or x_new [ii, jj] > col -1 :      continue    if y_new [ii, jj] < 0 or y_new [ii, jj] > row -1 :      continue    img_out[ii, jj, :] = img[new_yy, new_xx, :]plt.figure (1)plt.title('www.vevb.com')plt.imshow (img)plt.axis('off')plt.figure (2)plt.title('www.vevb.com')plt.imshow (img_out)plt.axis('off')plt.show()