import osprint(os.getcwd())

git clone https://github.com/fizyr/keras-retinanet.git%cd keras-retinanet/

!pip install .

!python setup.py build_ext --inplace

import numpy as np

import shutil

import pandas as pd

import os, sys, random

import xml.etree.ElementTree as ET

import pandas as pd

from os import listdir

from os.path import isfile, join

import matplotlib.pyplot as plt

from PIL import Image

import requests

import urllib

from keras_retinanet.utils.visualization import draw_box, draw_caption , label_color

from keras_retinanet.utils.image import preprocess_image, resize_image

pngPath='C:/Users/PraveenKumar/RetinaNet//maskDetectorJPEGImages/'

annotPath='C:/Users/PraveenKumar/RetinaNet//maskDetectorXMLfiles/'

data=pd.DataFrame(columns=['fileName','xmin','ymin','xmax','ymax','class'])

os.getcwd()#read All filesallfiles = [f for f in listdir(annotPath) if isfile(join(annotPath, f))

#Read all pdf files in images and then in text and store that in temp folder #Read all pdf files in images and then in text and store that in temp folder

for file in allfiles:

#print(file)

if (file.split(".")[1]=='xml'):

fileName='C:/Users/PraveenKumar/RetinaNet/maskDetectorJPEGImages/'+file.replace(".xml",'.jpg')

        tree = ET.parse(annotPath+file)

        root = tree.getroot()

        for obj in root.iter('object'):

            cls_name = obj.find('name').text

            xml_box = obj.find('bndbox')

            xmin = xml_box.find('xmin').text

            ymin = xml_box.find('ymin').text

            xmax = xml_box.find('xmax').text

            ymax = xml_box.find('ymax').text

            # Append rows in Empty Dataframe by adding dictionaries

            data = data.append({'fileName': fileName, 'xmin': xmin, 'ymin':ymin,'xmax':xmax,'ymax':ymax,'class':cls_name}, ignore_index=True)

# pick a random image  filepath = df.sample()['fileName'].values[0]

  # get all rows for this image  df2 = df[df['fileName'] == filepath]

  im = np.array(Image.open(filepath))

  # if there's a PNG it will have alpha channel  im = im[:,:,:3]

  for idx, row in df2.iterrows():

    box = [

      row['xmin'],

      row['ymin'],

      row['xmax'],

      row['ymax'],

    ]

    print(box)

    draw_box(im, box, color=(255, 0, 0))

  plt.axis('off')

  plt.imshow(im)

  plt.show()                        

show_image_with_boxes(data)

#Define labels & write them in a fileclasses = ['mask','noMask']with open('../maskDetectorClasses.csv', 'w') as f:

  for i, class_name in enumerate(classes):

    f.write(f'{class_name},{i}\n')         

if not os.path.exists('snapshots'):

  os.mkdir('snapshots')

python keras_retinanet/bin/train.py --freeze-backbone

            --random-transform \

            --weights {PRETRAINED_MODEL}

            --batch-size 8

            --steps 500

            --epochs 15  

            csv maskDetectorData.csv maskDetectorClasses.csv

from glob import glob

model_paths = glob('snapshots/resnet50_csv_0*.h5')

latest_path = sorted(model_paths)[-1]

print("path:", latest_path)from keras_retinanet import models

model = models.load_model(latest_path, backbone_name='resnet50')

model = models.convert_model(model)

label_map = {}for line in open('../maskDetectorClasses.csv'):

  row = line.rstrip().split(',')

  label_map[int(row[1])] = row[0]

#Write a function to choose one image randomly from your dataset and predict using Trained model.def show_image_with_predictions(df, threshold=0.6):

  # choose a random image  row = df.sample()

  filepath = row['fileName'].values[0]

  print("filepath:", filepath)

  # get all rows for this image  df2 = df[df['fileName'] == filepath]

  im = np.array(Image.open(filepath))

  print("im.shape:", im.shape)

  # if there's a PNG it will have alpha channel  im = im[:,:,:3]

  # plot true boxes  for idx, row in df2.iterrows():

    box = [

      row['xmin'],

      row['ymin'],

      row['xmax'],

      row['ymax'],

    ]

    print(box)

    draw_box(im, box, color=(255, 0, 0))

  ### plot predictions ###  # get predictions  imp = preprocess_image(im)

  imp, scale = resize_image(im)

  boxes, scores, labels = model.predict_on_batch(

    np.expand_dims(imp, axis=0)

  )

  # standardize box coordinates  boxes /= scale

  # loop through each prediction for the input image  for box, score, label in zip(boxes[0], scores[0], labels[0]):

    # scores are sorted so we can quit as soon    # as we see a score below threshold    if score < threshold:

      break    box = box.astype(np.int32)

    color = label_color(label)

    draw_box(im, box, color=color)

    class_name = label_map    caption = f"{class_name} {score:.3f}"

    draw_caption(im, box, caption)

    score, label=score, label

  plt.axis('off')

  plt.imshow(im)

  plt.show()

  return score, label

plt.rcParams['figure.figsize'] = [20, 10]