find_wally_pretty.py

from matplotlib import pyplot as plt
import numpy as np
import sys
import tensorflow as tf
import matplotlib
from PIL import Image
import matplotlib.patches as patches

model_path = './trained_model/frozen_inference_graph.pb'
image_path='eval_images/1.jpg'

def draw_box(box, image_np):
    #expand the box by 50%
    box += np.array([-(box[2] - box[0])/2, -(box[3] - box[1])/2, (box[2] - box[0])/2, (box[3] - box[1])/2]) 

    fig = plt.figure()
    ax = plt.Axes(fig, [0., 0., 1., 1.])
    fig.add_axes(ax)

    #draw blurred boxes around box
    ax.add_patch(patches.Rectangle((0,0),box[1]*image_np.shape[1], image_np.shape[0],linewidth=0,edgecolor='none',facecolor='w',alpha=0.8))
    ax.add_patch(patches.Rectangle((box[3]*image_np.shape[1],0),image_np.shape[1], image_np.shape[0],linewidth=0,edgecolor='none',facecolor='w',alpha=0.8))
    ax.add_patch(patches.Rectangle((box[1]*image_np.shape[1],0),(box[3]-box[1])*image_np.shape[1], box[0]*image_np.shape[0],linewidth=0,edgecolor='none',facecolor='w',alpha=0.8))
    ax.add_patch(patches.Rectangle((box[1]*image_np.shape[1],box[2]*image_np.shape[0]),(box[3]-box[1])*image_np.shape[1], image_np.shape[0],linewidth=0,edgecolor='none',facecolor='w',alpha=0.8))

    return fig, ax

detection_graph = tf.Graph()
with detection_graph.as_default():
    od_graph_def = tf.GraphDef()
    with tf.gfile.GFile(model_path, 'rb') as fid:
        serialized_graph = fid.read()
        od_graph_def.ParseFromString(serialized_graph)
        tf.import_graph_def(od_graph_def, name='')

def load_image_into_numpy_array(image):
  (im_width, im_height) = image.size
  return np.array(image.getdata()).reshape(
      (im_height, im_width, 3)).astype(np.uint8)


with detection_graph.as_default():
  with tf.Session(graph=detection_graph) as sess:
    image_np = load_image_into_numpy_array(Image.open(image_path))
    image_tensor = detection_graph.get_tensor_by_name('image_tensor:0')
    boxes = detection_graph.get_tensor_by_name('detection_boxes:0')
    scores = detection_graph.get_tensor_by_name('detection_scores:0')
    classes = detection_graph.get_tensor_by_name('detection_classes:0')
    num_detections = detection_graph.get_tensor_by_name('num_detections:0')
    # Actual detection.
    (boxes, scores, classes, num_detections) = sess.run(
        [boxes, scores, classes, num_detections],
        feed_dict={image_tensor: np.expand_dims(image_np, axis=0)})

    if scores[0][0] < 0.1:
        sys.exit('Wally not found :(')

    print('Wally found')
    fig, ax = draw_box(boxes[0][0], image_np)
  
    ax.imshow(image_np)
    plt.show()