def testImage(self):
print "Testing!"
self.queue.put("Working...")
nImage = image.resize((200, 150), Image.NEAREST)
imgArr = numpy.asarray(nImage)
t = time.time()
j = numpy.array(ExtractCharacters.LabelPixels(imgArr),
dtype=numpy.uint8)
print time.time() - t
imagesWithCenters = ExtractCharacters.PasteCharacters(imgArr, j)
numbers = []
if alreadyWorking.locked():
alreadyWorking.release()
for imageWCenter in imagesWithCenters:
testVec = map(norm, convert.getFeatureVecFromImg(imageWCenter[0]))
if alreadyWorking.locked():
return #another thread has come into play!
with lock:
net.ClearNodes()
net.SetInputs(testVec)
rawOutputs = net.ComputeOutput()
outputString = str(rawOutputs.index(max(rawOutputs)))
numbers.append((imageWCenter[1], outputString))
numbers.sort()
numbers = "".join(map(lambda elem: elem[1], numbers))
print numbers
self.queue.put("Neural network read: " + numbers)
print "done testing"
def testImage(self):
print "Testing!"
self.queue.put("Working...")
nImage = image.resize((200,150),Image.NEAREST)
imgArr = numpy.asarray(nImage)
t = time.time()
j = numpy.array(ExtractCharacters.LabelPixels(imgArr),dtype=numpy.uint8)
print time.time()-t
imagesWithCenters = ExtractCharacters.PasteCharacters(imgArr,j)
numbers = []
if alreadyWorking.locked():
alreadyWorking.release()
for imageWCenter in imagesWithCenters:
testVec = map(norm,convert.getFeatureVecFromImg(imageWCenter[0]))
if alreadyWorking.locked():
return #another thread has come into play!
with lock:
net.ClearNodes()
net.SetInputs(testVec)
rawOutputs = net.ComputeOutput()
outputString = str(rawOutputs.index(max(rawOutputs)))
numbers.append((imageWCenter[1],outputString))
numbers.sort()
numbers = "".join(map(lambda elem: elem[1], numbers))
print numbers
self.queue.put("Neural network read: " + numbers)
print "done testing"
def TestNumber(testVec):
testVec = map(norm,convert.getFeatureVecFromImg(testVec))
net.ClearNodes()
net.SetInputs(testVec)
rawOutputs = net.ComputeOutput()
outputs = map (Indicator,net.ComputeOutput())
print "Confidence Values:"
print ("\n" + fspace).join(textwrap.wrap(fspace+str(rawOutputs).ljust(10)))
print "Greatest Confidence:"
print fspace + str(rawOutputs.index(max(rawOutputs))) + " with " + str(100*max(rawOutputs)) + "% confidence"
print
return rawOutputs.index(max(rawOutputs))
def testDigit(self):
print "Testing!"
try:
featureVec = map(norm, convert.getFeatureVecFromImg(image))
except:
return
with lock:
net.ClearNodes()
net.SetInputs(featureVec)
rawOutputs = net.ComputeOutput()
outputs = map(Indicator, rawOutputs)
roundedOutputs = map(lambda x: "%.3f" % round(x, 3), rawOutputs)
outputString = "Confidence Values:\n" + \
tabspace+str(roundedOutputs).ljust(10) + "\n" + \
"Greatest Confidence:\n" + \
tabspace + str(rawOutputs.index(max(rawOutputs))) + " with %3.3f" % (100*max(rawOutputs)) + "% confidence"
self.queue.put(outputString)
print "done testing"
def testDigit(self):
print "Testing!"
try:
featureVec = map(norm,convert.getFeatureVecFromImg(image))
except:
return
with lock:
net.ClearNodes()
net.SetInputs(featureVec)
rawOutputs = net.ComputeOutput()
outputs = map (Indicator,rawOutputs)
roundedOutputs = map (lambda x: "%.3f" % round (x,3),rawOutputs)
outputString = "Confidence Values:\n" + \
tabspace+str(roundedOutputs).ljust(10) + "\n" + \
"Greatest Confidence:\n" + \
tabspace + str(rawOutputs.index(max(rawOutputs))) + " with %3.3f" % (100*max(rawOutputs)) + "% confidence"
self.queue.put(outputString)
print "done testing"