def process_data(r0):
r0 = lm.shuffle_points(r0)
### CPD registration
rtemp0, n, i = lm.get_shape_with_most_points(r0)
r1 = lm.register_cpd(r0, rtemp0)
### reorder points:
r2 = lm.reorder_points(r1, optimum_order=True, ensure_clockwise=True)
### correspondence:
rtemp1 = r2[i]
r3 = lm.set_npoints(r2, n)
r4 = lm.corresp_roll(r3, rtemp1)
return r4, rtemp1
#(0) Load data:
dirREPO = lm.get_repository_path()
names = [
'Bell', 'Comma', 'Device8', 'Face', 'Flatfish', 'Hammer', 'Heart',
'Horseshoe', 'Key'
]
name = names[0]
fnameCSV = os.path.join(dirREPO, 'Data', name, 'contours_sroc.csv')
r = lm.read_csv(fnameCSV)
i0, i1 = 1, 2
r0, r1 = r[i0], r[i1]
#(1) Process:
np.random.seed(0)
### shuffle:
rA0, rA1 = lm.shuffle_points(r0), lm.shuffle_points(r1)
### reorder:
rB0, rB1 = [
lm.reorder_points(r, optimum_order=True, ensure_clockwise=True)
for r in [rA0, rA1]
]
### optimum roll correspondence:
rD0 = rB0.copy()
rD1 = lm.corresp_roll(rB1, rB0)
### intermediary roll:
rC0 = rB0.copy()
rC1 = np.roll(rB1, 75, axis=0)
#(1) Plot:
plt.close('all')
lm.set_matplotlib_rcparams()
r3 = lm.set_npoints(r2, n)
r4 = lm.corresp_roll(r3, rtemp1)
return r4, rtemp1
#(0) Process data:
### load data
dirREPO = lm.get_repository_path()
names = [
'Bell', 'Comma', 'Device8', 'Face', 'Flatfish', 'Hammer', 'Heart',
'Horseshoe', 'Key'
]
name = names[6]
fname0 = os.path.join(dirREPO, 'Data', name, 'contours.csv')
r0 = lm.read_csv(fname0)
r0 = lm.shuffle_points(r0)
### CPD registration
rtemp0, n, i = lm.get_shape_with_most_points(r0)
r1 = lm.register_cpd(r0, rtemp0)
### reorder points:
r2 = lm.reorder_points(r1, optimum_order=True, ensure_clockwise=True)
### correspondence:
rtemp1 = r2[i]
r3 = lm.set_npoints(r2, n)
r4 = lm.corresp_roll(r3, rtemp1)
### process data:
np.random.seed(0) # Case 1
r1, rtemp1 = process_data(r0)
np.random.seed(2) # Case 2
r2, rtemp2 = process_data(r0)
### hypothesis tests: