myImageLibrary.py

import numpy as np
import scipy.misc
import cv2
import glob
import os
import math
import pickle
import base64


def showimage(title,img):
    cv2.imshow(title,img)
    cv2.waitKey(0)
    cv2.destroyAllWindows()
    
def extract_face(img,FACE_CASCADE,return_rectangles = False):
    img_gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
    rectangles = FACE_CASCADE.detectMultiScale(img_gray, 1.3, 5)
    faces = []
    for (x,y,w,h) in rectangles:
        cropped_face = img[y:y+h,x:x+w]
        faces.append(cropped_face)
    if return_rectangles:
        return faces, rectangles
    else:
        return faces

def extract_faces_bulk(img_list,FACE_CASCADE):
    return_list = []
    for img in img_list:
        try:
            faces = extract_face(img,FACE_CASCADE)
            if len(faces)>0:
                return_list.append(faces[0])
        except:
            pass
    return return_list

def crop_center(image,cropx,cropy):
    w = image.shape[1]
    h = image.shape[0]
    startx = w//2-(cropx//2)
    starty = h//2-(cropy//2)
    return image[starty:starty+cropy, startx:startx+cropx]

def resize_crop(image,square_size):
    w = image.shape[1]
    h = image.shape[0]
    min_dim = min(w,h)  
    max_square_image = crop_center(image, min_dim, min_dim)
    result = cv2.resize(max_square_image,(square_size,square_size),0,0)
    return result



def get_images(parent_directory):
    results = glob.glob(os.path.join(parent_directory,'*.jpg'))+glob.glob(os.path.join(parent_directory,'*.jpeg'))+glob.glob(os.path.join(parent_directory,'*.png'))
    images = []
    for result in results:
        images.append(cv2.imread(result))
    return images

def preprocess(img,transpose=True,normalize=False):
    #takes an image (face) of shape (x,x,3), transforms it to shape (1,3,x,x), pickles it and encodes to base64.
    #also normalizes before encoding if normalize=True
    #only transposes if tranpose=True
    if transpose:
        img = img.transpose(2,0,1)
    if normalize:
        img = np.around(img/255.0, decimals = 12) 
    img = np.expand_dims(img,axis=0)
    img_pkl = pickle.dumps(img)
    img_base64 = base64.b64encode(img_pkl)
    return img_base64