Adding Python script for Burton Machado

python/README

@@ -0,0 +1 @@
+Name your scripts <id>.py eg. 628.py. Your id is the same as you used in the legend.csv.

python/bmachado.py

@@ -0,0 +1,99 @@
+#http://docs.opencv.org/3.0-beta/doc/py_tutorials/py_imgproc/py_geometric_transformations/py_geometric_transformations.html#geometric-transformations
+
+import cv2
+import numpy as np
+import matplotlib.pyplot as plt
+import os
+from os import listdir
+from os.path import isfile, join
+
+#Image rotation.
+def rotate_transform(image, degrees):
+    img = cv2.imread(image, 0)
+    rows,cols = img.shape
+    M = cv2.getRotationMatrix2D((cols/2,rows/2),degrees,1)
+    dst = cv2.warpAffine(img,M,(cols,rows))
+    return (dst)
+
+#Affine transformation.
+def affine_transform(image, points1, points2):
+    img = cv2.imread(image)
+    rows, cols, ch = img.shape
+    pts1 = np.float32(points1) #upper-left, upper-right, bottom-left
+    pts2 = np.float32([points2]) # New locations
+    M = cv2.getAffineTransform(pts1, pts2)
+    dst = cv2.warpAffine(img, M, (cols, rows))
+    return(dst)
+
+#Perspective transformation.
+def perspective_transform(image, points1, points2):
+    img = cv2.imread(image)
+    rows, cols, ch = img.shape
+    pts1 = np.float32(points1)
+    pts2 = np.float32(points2)
+    M = cv2.getPerspectiveTransform(points1, points2)
+    dst = cv2.warpPerspective(img, M, (350, 350))
+    return (dst)
+
+# The main function to transform images. Pass a list of image files to this function.
+# This function should be run from the the folder containing the images.
+def transformImages(imageList):
+    for file in imageList:
+        filename, file_extension = os.path.splitext(file)
+        print('Processing image : ' + file)
+        if file.startswith('.') or file.endswith('.py') or file.endswith('.DS_Store'):
+            continue
+
+        dst = rotate_transform(file, 45)
+        cv2.imwrite(filename + '_rot_45' + file_extension, dst)
+
+        dst = rotate_transform(file, 90)
+        cv2.imwrite(filename + '_rot_90' + file_extension, dst)
+
+        dst = rotate_transform(file, 135)
+        cv2.imwrite(filename + '_rot_135' + file_extension, dst)
+
+        dst = rotate_transform(file, 180)
+        cv2.imwrite(filename + '_rot_180' + file_extension, dst)
+
+        dst = rotate_transform(file, 225)
+        cv2.imwrite(filename + '_rot_225' + file_extension, dst)
+
+        dst = rotate_transform(file, 270)
+        cv2.imwrite(filename + '_rot_270' + file_extension, dst)
+
+        dst = rotate_transform(file, 315)
+        cv2.imwrite(filename + '_rot_315' + file_extension, dst)
+
+        points1 = [[10, 10], [80, 20], [5, 80]]
+        points2 = [[5, 20], [80, 20], [15, 90]]
+        dst = affine_transform(file, points1, points2)
+        cv2.imwrite(filename + '_aff_1' + file_extension, dst)
+
+        points1 = [[5, 5], [80, 20], [10, 80]]
+        points2 = [[10, 20], [80, 20], [20, 90]]
+        dst = affine_transform(file, points1, points2)
+        cv2.imwrite(filename + '_aff_2' + file_extension, dst)
+
+        points1 = np.float32([[10, 10], [100, 10], [10, 100], [100, 100]])
+        points2 = np.float32([[0, 0], [100, 0], [0, 100], [100, 100]])
+        dst = perspective_transform(file, points1, points2)
+        cv2.imwrite(filename + '_pers_1' + file_extension, dst)
+
+# Start the image transformations.
+
+# Following is an example of how to use this code.
+# Create a list of image files.
+imagesDir = '/Users/burton/000-Semester_06_CUNY/622_Machine_Learning/Week_10/resultdir/'
+imageList = []
+allFiles = [f for f in listdir(imagesDir) if isfile(join('', f))]
+for file in allFiles:
+    imageList.append(imagesDir + file)
+
+# Pass the image list to transformImages function. The resulting iamges will be written to the folder from which the images were read.
+#This function takes the list of image files, transforms the images and writes them to the same folder as the image.
+#Please pass the full image file path to this function if running from a folder other than the folder containing the images.
+transformImages(imageList)
+
+#After running the above function you should see 10 transformations of each image in the folder containing that image.
+