def img_classify(self, img):
kernel = np.ones((5, 5), np.uint8)
thresh = l32.peakPick(img)
img_bin = abs(l32.threshold(img, 127) -
1) # Given this set of data we need to flip the values
img_props = np.array([l42.getproperties(img_bin)], dtype=np.float32)
prediction = self.svm.predict(img_props)
return prediction
def addTrainSamples(self, folder):
kernel = np.ones((5, 5), np.uint8)
class_key = 0
for c in self.classes:
for i in range(1, c[1] + 1):
filename = folder + '/' + c[0] + str(i) + '.jpg'
print(filename)
img = l25.loadImage(filename, 0)
thresh = l32.peakPick(img)
img_bin = abs(
l32.threshold(img, 127) -
1) # Given this set of data we need to flip the values
# plt.imshow(img_bin, cmap='hot', interpolation='nearest')
# plt.show()
img_props = np.array([l42.getproperties(img_bin)],
dtype=np.float32)
if self.samples.size == 0:
self.samples = img_props
else:
self.samples = np.append(self.samples, img_props, axis=0)
self.samples_labels = np.append(
self.samples_labels, np.array([class_key], dtype=np.int))
class_key += 1
return