featureGeneration.py

import readFile
import cv2
from skimage.feature import local_binary_pattern
import labels
import numpy as np
from scipy.stats import itemfreq 
from sklearn.preprocessing import normalize

folder="D:\data-1"
data=labels.test_image
images=readFile.resize_images(folder,data)
gray_images=readFile.grayscale_images(folder,data)
lbp_radius=2
lbp_numpoints=12
lbp_bin=2**lbp_numpoints

eps=1e-7

def calculate_lbp(images):
	lbp_features=[]
	for im in images:
		lbp = local_binary_pattern(im, lbp_numpoints, lbp_radius, method='default')
		(hist, hist_len) = np.histogram(lbp.ravel(),bins=np.arange(0, lbp_bin))
		hist = hist.astype("float")
		# hist /= hist.sum()
		lbp_features.append(hist)
	lbp_feat=np.array(lbp_features)
	return lbp_feat

def calculate_hog(images):
	winSize = (128,64)
	blockSize = (16,16)
	blockStride = (8,8)
	cellSize = (8,8)
	nbins = 9
	derivAperture = 1
	winSigma = 4.
	histogramNormType = 0
	L2HysThreshold = 2.0000000000000001e-01
	gammaCorrection = 0
	nlevels = 64
	winStride = (8,8)
	padding = (8,8)
	hog_features=[]
	for im in images:
		hog=cv2.HOGDescriptor(winSize,blockSize,blockStride,cellSize,nbins,derivAperture,winSigma,histogramNormType,L2HysThreshold,gammaCorrection,nlevels)
		hist=hog.compute(im)
		hog_features.append(hist)
	return hog_features

# def calculate_orb(images):
# 	orb = cv2.ORB_create()

# 	orb_features=[]
# 	for img in images:
# 		kp = orb.detect(img,None)
# 		kp, des = orb.compute(img, kp)
# 		print des
# 		break
	
# lbp_feat=calculate_lbp(gray_images)
# print lbp_feat

# print lbp_feat.shape

# hog_feat=calculate_hog()
# print hog_feat[0]

# calculate_orb(gray_images)