before publishing the code

3D_show.py

@@ -0,0 +1,102 @@
+# demonstrate the 3D models
+import numpy as np
+import pickle
+import json
+import os, glob
+import math
+
+import matplotlib.pyplot as plt
+from mpl_toolkits.mplot3d import Axes3D
+from matplotlib import cm
+from matplotlib.ticker import LinearLocator, FormatStrFormatter
+import matplotlib.pyplot as plt
+from mpl_toolkits.mplot3d import Axes3D
+import pdb
+
+def valid_region_I(cfg, I):
+	resolution = [[cfg[i]['szy_sensor'], cfg[i]['szx_sensor']] for i in range(len(cfg))]
+	rows_cut = int(\
+		((cfg[-1]['gauss'].shape[0]-1)/2+\
+		(cfg[-1]['ext_f'].shape[1]-1)/2)*\
+		resolution[0][0]/resolution[-1][0]
+	)
+	cols_cut = int(\
+		((cfg[-1]['gauss'].shape[1]-1)/2+\
+		(cfg[-1]['ext_f'].shape[2]-1)/2)*\
+		resolution[0][1]/resolution[-1][1]
+	)
+
+	rows = cfg[0]['szx_sensor']
+	cols = cfg[0]['szy_sensor']
+
+	I = \
+		I[
+			cols_cut:cols-cols_cut,
+			rows_cut:rows-rows_cut
+		]
+	return I
+
+# depth_file = "./models/"+\
+# 		"121.pickle" #74 start of shield #88 start of take 2 with shield
+depth_file = "./models/quad/"+\
+		"125.pickle"
+
+with open(depth_file,'rb') as f:
+	depth_data = pickle.load(f)
+
+x = np.arange(depth_data['Zf'].shape[1])
+y = np.arange(depth_data['Zf'].shape[0])
+[X,Y] = np.meshgrid(x,y)
+
+# read data
+I_0 = depth_data['I_0']
+I_1 = depth_data['I_1']
+Z = depth_data['Zf']
+conf = depth_data['conff']
+
+# set the depth range
+dmin = -2.0
+dmax = -0.8
+Z[np.where(Z <= dmin)] = np.NaN
+Z[np.where(Z >= dmax)] = np.NaN
+
+
+
+cfg_file = "./opt_results/pyConfLensFlowNetFast_ext/"+\
+		"1x1t-text34-py4-setup5-one-sequential-regularize-nothreshold.pickle"
+with open(cfg_file,'rb') as f:
+	cfg_data = pickle.load(f)
+cfg = cfg_data['cfg']
+I_0 = valid_region_I(cfg[0:-1], I_0)
+I_1 = valid_region_I(cfg[0:-1], I_1)
+
+# draw the captured image
+fig = plt.figure()
+ax = fig.add_subplot(1,3,1, title="First image")
+plt.imshow(I_0, cmap = 'gray',vmin=0, vmax=255)
+plt.axis('off')
+ax = fig.add_subplot(1,3,2, title="Second image")
+plt.imshow(I_1, cmap = 'gray',vmin=0, vmax=255)
+plt.axis('off')
+# plt.imshow(conf, cmap= 'gray')
+
+# confidence threshold
+conf_thre = [0.999]
+for i in range(len(conf_thre)):
+	Z_cut = Z
+	Z_cut[np.where(conf<= conf_thre[i])] = np.NaN
+	# Z_cut[np.where(Z<=dmin)] = np.NaN
+	# Z_cut[np.where(Z>=dmax)] = np.NaN
+
+	ax = fig.add_subplot(1,3,i+3, title="Depth map, conf > "+str(conf_thre[i]), projection='3d')
+	Z_cut = Z_cut[::-1]
+	# plt.imshow(Z_cut, cmap='jet', origin='low', vmin=dmin, vmax=dmax)
+	Z_cut[np.where(Z_cut>dmax)] = np.NaN
+	Z_cut[np.where(Z_cut<dmin)] = np.NaN
+	xx,yy = np.meshgrid(np.arange(Z_cut.shape[1]),np.arange(Z_cut.shape[0]))
+	pdb.set_trace()
+	ax.plot_surface(xx,yy,Z_cut)
+	# plt.colorbar()
+	plt.axis('off')
+plt.show()
+

3D_show_layer.py

@@ -0,0 +1,115 @@
+# demonstrate the 3D models
+import numpy as np
+import pickle
+import json
+import os, glob
+import math
+
+import matplotlib.pyplot as plt
+from mpl_toolkits.mplot3d import Axes3D
+from matplotlib import cm
+from matplotlib.ticker import LinearLocator, FormatStrFormatter
+import pdb
+
+def valid_region_I(cfg, I):
+	resolution = [[cfg[i]['szy_sensor'], cfg[i]['szx_sensor']] for i in range(len(cfg))]
+	rows_cut = int(\
+		((cfg[-1]['gauss'].shape[0]-1)/2+\
+		(cfg[-1]['ext_f'].shape[1]-1)/2)*\
+		resolution[0][0]/resolution[-1][0]
+	)
+	cols_cut = int(\
+		((cfg[-1]['gauss'].shape[1]-1)/2+\
+		(cfg[-1]['ext_f'].shape[2]-1)/2)*\
+		resolution[0][1]/resolution[-1][1]
+	)
+
+	rows = cfg[0]['szx_sensor']
+	cols = cfg[0]['szy_sensor']
+
+	I = \
+		I[
+			cols_cut:cols-cols_cut,
+			rows_cut:rows-rows_cut
+		]
+	return I
+
+depth_file = "./models/"+\
+		"leopard.pickle"
+with open(depth_file,'rb') as f:
+	depth_data = pickle.load(f)
+
+x = np.arange(depth_data['Z'].shape[1])
+y = np.arange(depth_data['Z'].shape[0])
+[X,Y] = np.meshgrid(x,y)
+
+# read data
+I_0 = depth_data['I_0']
+I_1 = depth_data['I_1']
+Z = depth_data['Z']
+conf = depth_data['conf']
+
+# set the depth range
+dmin = -0.88
+dmax = -0.48
+Z[np.where(Z <= dmin)] = np.NaN
+Z[np.where(Z >= dmax)] = np.NaN
+
+cfg_file = "./opt_results/pyConfLensFlowNetFast_ext/"+\
+		"1x1t-text34-py4-setup5-one-sequential-regularize-nothreshold.pickle"
+with open(cfg_file,'rb') as f:
+	cfg_data = pickle.load(f)
+cfg = cfg_data['cfg']
+I_0 = valid_region_I(cfg, I_0)
+I_1 = valid_region_I(cfg, I_1)
+
+# confidence threshold
+conf_thre = 0.99
+
+# draw the captured image
+fig = plt.figure()
+
+ax = fig.add_subplot(5,3,1, title="First image")
+plt.imshow(I_0, cmap = 'gray')
+plt.axis('off')
+
+ax = fig.add_subplot(5,3,2, title="Second image")
+plt.imshow(I_1, cmap = 'gray')
+plt.axis('off')
+
+ax = fig.add_subplot(5,3,3, title="$Z$")
+Z_cut = Z
+Z_cut[np.where(conf<= conf_thre)] = np.NaN
+Z_cut = Z_cut[::-1]
+plt.imshow(Z_cut, interpolation='bilinear', origin='low')
+plt.colorbar()
+plt.axis('off')
+
+fig_name = [
+	"$Z^{0,0,0}$",
+	"$Z^{0,1,0}$",
+	"$Z^{0,0,1}$",
+	"$Z^{1,0,0}$",
+	"$Z^{1,1,0}$",
+	"$Z^{1,0,1}$",
+	"$Z^{2,0,0}$",
+	"$Z^{2,1,0}$",
+	"$Z^{2,0,1}$",
+	"$Z^{3,0,0}$",
+	"$Z^{3,1,0}$",
+	"$Z^{3,0,1}$",
+]
+for i in range(12):
+	ax = fig.add_subplot(5,3,i+4, title=fig_name[i])
+	Z_cut = depth_data['Z'+str(i)]
+	conf = depth_data['conf'+str(i)]
+	Z_cut[np.where(conf<= conf_thre)] = np.NaN
+	Z_cut[np.where(Z_cut <= dmin)] = np.NaN
+	Z_cut[np.where(Z_cut >= dmax)] = np.NaN
+
+	Z_cut = Z_cut[::-1]
+	plt.imshow(Z_cut, interpolation='bilinear', origin='low')
+	plt.colorbar()
+	plt.axis('off')
+plt.show()
+

arrangeCUReT.m

@@ -0,0 +1,17 @@
+% Arrange the CUReT dataset so that all textures are arranged from 0 to
+% 5611
+infolder = 'curetgrey';
+outfolder = 'Qi_curetgrey';
+samples = dir(infolder);
+samples(1:2) = [];
+num = 1;
+idx = 0;
+for i = 1:length(samples)
+    images = dir(fullfile(infolder,samples(i).name,'*.png'));
+    for j = 1:length(images)
+        img = imread(fullfile(infolder,samples(i).name,images(j).name));
+        imwrite(img, fullfile(outfolder,sprintf('%04d.png',idx)));
+        idx = idx + 1;
+    end
+    num = num + length(images);
+end