read_images.py

# -*- coding: utf-8 -*-
"""
Created on Mon Mar 11 22:44:41 2019

@author: Gireesh Sundaram
"""

import cv2
import numpy as np
import math
import pandas as pd
import codecs

from configuration import window_height, window_width, window_shift, MPoolLayers_H

#%%
vec_per_window = window_width / math.pow(2, MPoolLayers_H)

#%%
#reading the class files
data = {}
with codecs.open("Data\\class.txt", 'r', encoding='utf-8') as cF:
    data = cF.read().split('\n')
    
#%%
def returnClasses(string):
    text = list(string)
    text = ["<SPACE>"] + ["<SPACE>" if x==" " else x for x in text] + ["<SPACE>"]
    classes = [data.index(x) if x in data else 2 for x in text]
    return classes
    
#%%
infile = pd.read_csv(".\\Data\\list.csv")

#%%
inputList = []
seqLens = []
targetList = []

for record in range(0, len(infile)):
    path = infile["Path"][record]
    annotation = infile["Annotation"][record]
    #reading images from the path
    print("Reading file: " + path)
    image = cv2.imread(path, cv2.IMREAD_GRAYSCALE)
    h, w = np.shape(image)
    print(np.shape(image))
    
    #setting factor if the image is greater than the window height
    if (h > window_height): factor = window_height/h
    else: factor = 1
    
    #resizing the image to the specified window height
    image = cv2.resize(image, None, fx=factor, fy=factor, interpolation = cv2.INTER_CUBIC)
    h, w = np.shape(image)
        
    featureSet = []
    
    winId = 0
    while True:
        s = winId * window_shift
        e = s + window_width
        
        if e > w:
            break
        
        wnd = image[:h, s:e]
        featureSet.append(wnd)
        winId = winId + 1
        
    inputList.append(featureSet)
    
    #taking the sequence length
    winId = 0
    wpd = 0
    while True:
        s = winId * window_shift
        e = s + window_width
        
        if e > w:
            sl = (winId+1) * vec_per_window
            seqLens.append(sl)
            break
        winId = winId + 1
        
    #taking the text annotations from the csv file and converting to labels
    targetList.append(returnClasses(annotation))