def Featurize_X_Histogram(x_raw):
x = []
for sample in x_raw:
# x-gradient histogram
image = Image.open(sample)
xGradients = Assignment5Support.Convolution3x3(image, [[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
xFeatures = CalculateHistogramFeatures(xGradients)
x.append(xFeatures)
return x
def Featurize_X_Gradient(x_raw):
x = []
for sample in x_raw:
# x-gradient 9 grids of 8x8 pixels
image = Image.open(sample)
xGradients = Assignment5Support.Convolution3x3(image, [[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
xFeatures = CalculateGradientFeatures(xGradients)
x.append(xFeatures)
return x
def Featurize_Y_Histogram(x_raw):
x = []
for sample in x_raw:
# y-gradient histogram
image = Image.open(sample)
yGradients = Assignment5Support.Convolution3x3(image, [[1, 2, 1],[0, 0, 0],[-1, -2, -1]])
yFeatures = CalculateHistogramFeatures(yGradients)
x.append(yFeatures)
return x
def Featurize_Y_Gradient(x_raw):
x = []
for sample in x_raw:
# y-gradient 9 grids of 8x8 pixels
image = Image.open(sample)
yGradients = Assignment5Support.Convolution3x3(image, [[1, 2, 1],[0, 0, 0],[-1, -2, -1]])
yFeatures = CalculateGradientFeatures(yGradients)
x.append(yFeatures)
return x
def Featurize(x_raw):
x = []
for sample in x_raw:
features = []
image = Image.open(sample)
yGradients = Assignment5Support.Convolution3x3(
image, [[1, 2, 1], [0, 0, 0], [-1, -2, -1]])
xGradients = Assignment5Support.Convolution3x3(
image, [[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
# y-gradient 9 grids of 8x8 pixels
features += CalculateGradientFeatures(yGradients)
# y-gradient histogram
features += CalculateHistogramFeatures(yGradients)
# x-gradient 9 grids of 8x8 pixels
features += CalculateGradientFeatures(xGradients)
# x-gradient histogram
features += CalculateHistogramFeatures(xGradients)
x.append(features)
return x
## Some of this references my answers to previous assignments.
## Replace my references with references to your answers to those assignments.
## IMPORTANT NOTE !!
## Remember to install the Pillow library (which is required to execute 'import PIL')
import Assignment5Support
import numpy as np
## NOTE update this with your equivalent code..
import TrainTestSplit
kDataPath = "/Users/Mims/Documents/school/assignments/BlinkSupport/dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = TrainTestSplit.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
print("Calculating features...")
(xTrain, xTest) = Assignment5Support.Featurize(xTrainRaw,
xTestRaw,
includeGradients=True,
includeRawPixels=False,
includeIntensities=True)
yTrain = yTrainRaw
## Replace my references with references to your answers to those assignments.
## IMPORTANT NOTE !!
## Remember to install the Pillow library (which is required to execute 'import PIL')
import matplotlib.pyplot as plt
import Assignment5Support
import Featurize
import RandomForest
import EvaluationsStub
if __name__ == "__main__":
kDataPath = "..\\..\\dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = Assignment5Support.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
yTrain = yTrainRaw
yTest = yTestRaw
############################
print("========== Use Random Forest for Y-Gradient ==========")
print("### Calculating features ###")
xTrain_Y_Gradient = Featurize.Featurize_Y_Gradient(xTrainRaw)
xTest_Y_Gradient = Featurize.Featurize_Y_Gradient(xTestRaw)
## Some of this references my answers to previous assignments.
## Replace my references with references to your answers to those assignments.
## IMPORTANT NOTE !!
## Remember to install the Pillow library (which is required to execute 'import PIL')
## Remember to install Pytorch: https://pytorch.org/get-started/locally/ (if you want GPU you need to figure out CUDA...)
import Assignment5Support
## NOTE update this with your equivalent code.
import TrainTestSplit
kDataPath = "..\\Datasets\\FaceData\\dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = TrainTestSplit.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
from PIL import Image
import torch
import torchvision.transforms as transforms
import CrossValidationSupport
torch.cuda.set_device(0)
import numpy as np
import copy
import CrossValidationSupport
import Evaluations
import ErrorBounds
import NeuralNetworkModel
predictionThreshold = .5
#thresholds = [.5]
#for i in range(0,101):
# thresholds.append(i*.01)
kDataPath = "..\\Datasets\\FaceData\\dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath, includeLeftEye = True, includeRightEye = True)
(xTrainRaw, yTrainRaw, xTestRaw, yTestRaw) = TrainTestSplit.TrainTestSplit(xRaw, yRaw, percentTest = .25)
print("Train is %f percent closed." % (sum(yTrainRaw)/len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw)/len(yTestRaw)))
print("Calculating features...")
(xTrain, xTest) = Assignment5Support.Featurize(xTrainRaw, xTestRaw, includeGradients=False, includeXGradients=False, includeYGradients=False, includeYHistogram=False, includeXHistogram=False, includeRawPixels=False, includeIntensities=True)
yTrain = yTrainRaw
yTest = yTestRaw
k = 5
UseCrossValidation = True
## Some of this references my answers to previous assignments.
## Replace my references with references to your answers to those assignments.
## IMPORTANT NOTE !!
## Remember to install the Pillow library (which is required to execute 'import PIL')
import Assignment5Support
import EvaluationsStub
import matplotlib.pyplot as plt
import numpy as np
if __name__ == "__main__":
kDataPath = "..\\..\\dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = Assignment5Support.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
print('Featurize the Samples')
(xTrain, xTest) = Assignment5Support.Featurize(xTrainRaw,
xTestRaw,
includeGradients=False,
includeRawPixels=False,
includeIntensities=True)
yTrain = np.array(yTrainRaw)[np.newaxis].T
yTest = np.array(yTestRaw)[np.newaxis].T
## Some of this references my answers to previous assignments.
## Replace my references with references to your answers to those assignments.
## IMPORTANT NOTE !!
## Remember to install the Pillow library (which is required to execute 'import PIL')
import Assignment5Support
import matplotlib.pyplot as plt
import numpy as np
if __name__ == "__main__":
kDataPath = "..\\..\\dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = Assignment5Support.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
print("Calculating features...")
(xTrain, xTest) = Assignment5Support.Featurize(xTrainRaw,
xTestRaw,
includeGradients=True,
includeRawPixels=False,
includeIntensities=False)
yTrain = yTrainRaw
yTest = yTestRaw
from PIL import Image
import Assignment5Support
import Featurize
image = Image.open(
'../dataset_B_Eye_Images/openRightEyes/Oscar_DLeon_0001_R.jpg')
# y-gradient 9 grids of 8x8 pixels
yGradients = Assignment5Support.Convolution3x3(
image, [[1, 2, 1], [0, 0, 0], [-1, -2, -1]])
yFeatures = Featurize.CalculateGradientFeatures(yGradients)
print(yFeatures[:12])
# x-gradient 9 grids of 8x8 pixels
xGradients = Assignment5Support.Convolution3x3(
image, [[-1, 0, 1], [-2, 0, 2], [-1, 0, 1]])
xFeatures = Featurize.CalculateGradientFeatures(xGradients)
print(xFeatures[:12])
# y-graident 5-bin histogram
yFeatures = Featurize.CalculateHistogramFeatures(yGradients)
print(yFeatures[:5])
# # x-gradient 5-bin histogram
xFeatures = Featurize.CalculateHistogramFeatures(xGradients)
print(xFeatures[:5])
# result:
# [-59, 63, -6.59375, -92, 78]
# [-125, 249, 13.3125, -196, 326]
# [0.6545138888888888, 0.2482638888888889, 0.06597222222222222, 0.024305555555555556, 0.006944444444444444]
import Assignment5Support
import numpy as np
import NeuralNetStack
import TrainTestSplit
import Evaluations
kDataPath = "/Users/Mims/Documents/school/assignments/BlinkSupport/dataset_B_Eye_Images"
(xRaw, yRaw) = Assignment5Support.LoadRawData(kDataPath,
includeLeftEye=True,
includeRightEye=True)
(xTrainRaw, yTrainRaw, xTestRaw,
yTestRaw) = TrainTestSplit.TrainTestSplit(xRaw, yRaw, percentTest=.25)
print("Train is %f percent closed." % (sum(yTrainRaw) / len(yTrainRaw)))
print("Test is %f percent closed." % (sum(yTestRaw) / len(yTestRaw)))
print("Calculating features...")
(xTrain, xTest) = Assignment5Support.Featurize(xTrainRaw,
xTestRaw,
includeGradients=False,
includeRawPixels=False,
includeIntensities=True)
yTrain = yTrainRaw
yTest = yTestRaw
#T = [(t0, [1,0,0,0,0]), (t1, [0,1,0,0,0]), (t2, [0,0,1,0,0]), (t3, [0,0,0,1,0]), (t4, [0,0,0,0,1])]
trainingData = []
for i in range(len(xTrain)):