def trainGest(gesture_name):
SAMPLES = 10
path = "Gestures/"
gesture = gesture_name;
print "--- training gesture " + gesture_name + " ---"
# Init output file
output = open(path + gesture + "/" + "out.txt", "w+")
output.write("---N0---N1---N2---N3---N4---N5---xs---ys---zs---")
output.write("\n")
# Init trained_gestures file
trained = open("trained.txt", "r+")
if not trained.read(1):
trained.close()
trained = open("trained.txt", "w+")
trained.write("- Name ----N0---N1---N2---N3---N4---N5---xs---ys---zs---")
trained.close()
trained = open("trained.txt", "a+")
trained.write("\n")
# Init dependent variables
sum_N0 = 0;
sum_N1 = 0;
sum_N2 = 0;
sum_N3 = 0;
sum_N4 = 0;
sum_N5 = 0;
sum_xs = 0;
sum_ys = 0;
sum_zs = 0;
min_N0 = 70000;
min_N1 = 70000;
min_N2 = 70000;
min_N3 = 70000;
min_N4 = 70000;
min_N5 = 70000;
min_xs = 70000;
min_ys = 70000;
min_zs = 70000;
max_N0 = -80000;
max_N1 = -80000;
max_N2 = -80000;
max_N3 = -80000;
max_N4 = -80000;
max_N5 = -80000;
max_xs = -80000;
max_ys = -80000;
max_zs = -80000;
k = 0
for i in range(0, SAMPLES):
print "Sensor "
sensorData = get_data()
print "SAMPLE" + str(i)
print sensorData[0]
print sensorData[1]
dataS0 = sensorData[0].split()
dataS1 = sensorData[1].split()
dataS2 = sensorData[2].split()
dataS3 = sensorData[3].split()
dataS4 = sensorData[4].split()
dataS5 = sensorData[5].split()
print dataS1[0]
x0 = float(dataS0[0])
y0 = float(dataS0[1])
z0 = float(dataS0[2])
x1 = float(dataS1[0])
y1 = float(dataS1[1])
z1 = float(dataS1[2])
x2 = float(dataS2[0])
y2 = float(dataS2[1])
z2 = float(dataS2[2])
x3 = float(dataS3[0])
y3 = float(dataS3[1])
z3 = float(dataS3[2])
x4 = float(dataS4[0])
y4 = float(dataS4[1])
z4 = float(dataS4[2])
x5 = float(dataS5[0])
y5 = float(dataS5[1])
z5 = float(dataS5[2])
# Calculate and print results
N0_out = N0(x0, y0, z0, 1, 1, 1)
N1_out = N1(x1, y1, z1, 1, 1, 1)
N2_out = N2(x2, y2, z2, 1, 1, 1)
N3_out = N3(x3, y3, z3, 1, 1, 1)
N4_out = N4(x4, y4, z4, 1, 1, 1)
N5_out = N5(x5, y5, z5, 1, 1, 1)
xs_out = xs(x0, x1, x2, x3, x4, x5, 1, 1, 1, 1, 1, 1)
ys_out = ys(y0, y1, y2, y3, y4, y5, 1, 1, 1, 1, 1, 1)
zs_out = zs(z0, z1, z2, z3, z4, z5, 1, 1, 1, 1, 1, 1)
print "N0:"
print N0_out
print "N1:"
print N1_out
print "N2:"
print N2_out
print "N3:"
print N3_out
print "N4:"
print N4_out
print "N5:"
print N5_out
print "xs:"
print xs_out
print "ys:"
print ys_out
print "zs:"
print zs_out
sum_N0 += N0_out
sum_N1 += N1_out
sum_N2 += N2_out
sum_N3 += N3_out
sum_N4 += N4_out
sum_N5 += N5_out
sum_xs += xs_out
sum_ys += ys_out
sum_zs += zs_out
if min_N0 > N0_out:
min_N0 = N0_out
if min_N1 > N1_out:
min_N1 = N1_out
if min_N2 > N2_out:
min_N2 = N2_out
if min_N3 > N3_out:
min_N3 = N3_out
if min_N4 > N4_out:
min_N4 = N4_out
if min_N5 > N5_out:
min_N5 = N5_out
if min_xs > xs_out:
min_xs = xs_out
if min_ys > ys_out:
min_ys = ys_out
if min_zs > zs_out:
min_zs = zs_out
print "max_n1 = " + str(max_N1)
print "n1_out= " + str(N1_out)
if max_N0 < N0_out:
max_N0 = N0_out
if max_N1 < N1_out:
max_N1 = N1_out
print "max_n1 = " + str(max_N1)
if max_N2 < N2_out:
max_N2 = N2_out
if max_N3 < N3_out:
max_N3 = N3_out
if max_N4 < N4_out:
max_N4 = N4_out
if max_N5 < N5_out:
max_N5 = N5_out
if max_xs < xs_out:
max_xs = xs_out
if max_ys < ys_out:
max_ys = ys_out
if max_zs < zs_out:
max_zs = zs_out
output.write(" ")
output.write("{} ".format((N0_out)))
output.write(" " + str(N1_out) + " ")
output.write(" " + str(N2_out) + " ")
output.write(" " + str(N3_out) + " ")
output.write(" " + str(N4_out) + " ")
output.write(" " + str(N5_out) + " ")
output.write(" " + str(xs_out) + " ")
output.write(" " + str(ys_out) + " ")
output.write(" " + str(zs_out) + " ")
output.write("\n")
# print k
k += 1
print "-----------------------"
print "Done collecting data"
k = k-1
med_N0 = float("{0:.2f}".format(sum_N0 / float(k)))
med_N1 = float("{0:.2f}".format(sum_N1 / float(k)))
med_N2 = float("{0:.2f}".format(sum_N2 / float(k)))
med_N3 = float("{0:.2f}".format(sum_N3 / float(k)))
med_N4 = float("{0:.2f}".format(sum_N4 / float(k)))
med_N5 = float("{0:.2f}".format(sum_N5 / float(k)))
med_xs = float("{0:.2f}".format(sum_xs / float(k)))
med_ys = float("{0:.2f}".format(sum_ys / float(k)))
med_zs = float("{0:.2f}".format(sum_zs / float(k)))
delta_N0 = max_N0 - min_N0
delta_N1 = max_N1 - min_N1
delta_N2 = max_N2 - min_N2
delta_N3 = max_N3 - min_N3
delta_N4 = max_N4 - min_N4
delta_N5 = max_N5 - min_N5
delta_xs = max_xs - min_xs
delta_ys = max_ys - min_ys
delta_zs = max_zs - min_zs
output.write("------------ average values ---------------\n")
output.write(" " + str(med_N0))
output.write(" " + str(med_N1))
output.write(" " + str(med_N2))
output.write(" " + str(med_N3))
output.write(" " + str(med_N4))
output.write(" " + str(med_N5))
output.write(" " + str(med_xs))
output.write(" " + str(med_ys))
output.write(" " + str(med_zs))
output.write("\n")
output.write("----------- delta ----------------\n")
output.write(" " + str(delta_N0))
output.write(" " + str(delta_N1) + " ")
output.write(" " + str(delta_N2) + " ")
output.write(" " + str(delta_N3) + " ")
output.write(" " + str(delta_N4) + " ")
output.write(" " + str(delta_N5) + " ")
output.write(" " + str(delta_xs) + " ")
output.write(" " + str(delta_ys) + " ")
output.write(" " + str(delta_zs) + " ")
output.write("\n")
# # write minimal values
# trained.write(gesture)
# trained.write(" ")
# trained.write(" " + str(min_N0))
# trained.write(" " + str(min_N1))
# trained.write(" " + str(min_N2))
# trained.write(" " + str(min_N3))
# trained.write(" " + str(min_N4))
# trained.write(" " + str(min_N5))
# trained.write(" " + str(min_xs))
# trained.write(" " + str(min_ys))
# trained.write(" " + str(min_zs) + "\n")
print "trained:" + str(med_N0) + " " + str(int(round(med_N1)))
# write average values
trained.write(gesture)
trained.write(" ")
trained.write(" " + str(int(round(med_N1))))
trained.write(" " + str(int(round(med_N1))))
trained.write(" " + str(int(round(med_N2))))
trained.write(" " + str(int(round(med_N3))))
trained.write(" " + str(int(round(med_N4))))
trained.write(" " + str(int(round(med_N5))))
trained.write(" " + str(int(round(med_xs))))
trained.write(" " + str(int(round(med_ys))))
trained.write(" " + str(int(round(med_zs))))
trained.write("\n\r")
# # write maximum values
# trained.write(gesture)
# trained.write(" ")
# trained.write(" " + str(max_N0))
# trained.write(" " + str(max_N1))
# trained.write(" " + str(max_N2))
# trained.write(" " + str(max_N3))
# trained.write(" " + str(max_N4))
# trained.write(" " + str(max_N5))
# trained.write(" " + str(max_xs))
# trained.write(" " + str(max_ys))
# trained.write(" " + str(max_zs) + "\n")
output.close()
trained.close()
def recogGest():
# global gesture_recog
gesture_recog ="None"
trained = open("trained.txt")
gestures = trained.readlines()
print "gestures: "
print gestures[1]
gestures.remove(gestures[0])
sensorData = get_data()
dataS0 = sensorData[0].split()
dataS1 = sensorData[1].split()
dataS2 = sensorData[2].split()
dataS3 = sensorData[3].split()
dataS4 = sensorData[4].split()
dataS5 = sensorData[5].split()
x0 = float(dataS0[0])
y0 = float(dataS0[1])
z0 = float(dataS0[2])
x1 = float(dataS1[0])
y1 = float(dataS1[1])
z1 = float(dataS1[2])
x2 = float(dataS2[0])
y2 = float(dataS2[1])
z2 = float(dataS2[2])
x3 = float(dataS3[0])
y3 = float(dataS3[1])
z3 = float(dataS3[2])
x4 = float(dataS4[0])
y4 = float(dataS4[1])
z4 = float(dataS4[2])
x5 = float(dataS5[0])
y5 = float(dataS5[1])
z5 = float(dataS5[2])
# calculate neuron values
N0_out = N0(x0, y0, z0, 1, 1, 1)
N1_out = N1(x1, y1, z1, 1, 1, 1)
N2_out = N2(x2, y2, z2, 1, 1, 1)
N3_out = N3(x3, y3, z3, 1, 1, 1)
N4_out = N4(x4, y4, z4, 1, 1, 1)
N5_out = N5(x5, y5, z5, 1, 1, 1)
xs_out = xs(x0, x1, x2, x3, x4, x5, 1, 1, 1, 1, 1, 1)
ys_out = ys(y0, y1, y2, y3, y4, y5, 1, 1, 1, 1, 1, 1)
zs_out = zs(z0, z1, z2, z3, z4, z5, 1, 1, 1, 1, 1, 1)
min_k = 70000
for gesture in gestures:
print gesture
values = gesture.replace("\n", "").replace("\r", "").split()
print "values:"
print values
if values != []:
# print "values:"
# print values
N0_current = values[1]
N1_current = values[2]
N2_current = values[3]
N3_current = values[4]
N4_current = values[5]
N5_current = values[6]
xs_current = values[7]
ys_current = values[8]
zs_current = values[9]
if N0_out == 0:
N0_out += 1
if N1_out == 0:
N1_out += 1
if N2_out == 0:
N2_out += 1
if N3_out == 0:
N3_out += 1
if N4_out == 0:
N4_out += 1
if N5_out == 0:
N5_out += 1
if xs_out == 0:
xs_out += 1
if ys_out == 0:
ys_out += 1
if zs_out == 0:
zs_out += 1
# similarity coefficient
k_N0 = abs(float(N0_current) - float(N0_out))
k_N1 = abs(float(N1_current) - float(N1_out))
k_N2 = abs(float(N2_current) - float(N2_out))
k_N3 = abs(float(N3_current) - float(N3_out))
k_N4 = abs(float(N4_current) - float(N4_out))
k_N5 = abs(float(N5_current) - float(N5_out))
k_xs = abs(float(xs_current) - float(xs_out))
k_ys = abs(float(ys_current) - float(ys_out))
k_zs = abs(float(zs_current) - float(zs_out))
k_sum = k_N0 + k_N1 + k_N2 + k_N3 + k_N4 + k_N5 + k_xs + k_ys + k_zs
# print k_N0, k_N1, k_N2, k_N3, k_N4, k_N5, k_xs, k_ys, k_zs, k_sum
diff_k = k_sum / 9
# print values[0], diff_k
if diff_k < min_k:
min_k = diff_k
gesture_recog = values[0]
return gesture_recog
