def main(unused_argv):
with tf.device("/cpu:0"):
image = load_image_from_path("images/image.jpg")[np.newaxis, ...]
print(image.dtype)
g = tf.Graph()
with g.as_default():
box_extractor = BoxExtractor(get_faster_rcnn_config())
inputs = tf.placeholder(tf.float32, shape=image.shape)
boxes, scores, cropped_inputs = box_extractor(inputs)
with tf.Session(graph=g) as sess:
saver = tf.train.Saver(var_list=box_extractor.variables)
saver.restore(sess, get_faster_rcnn_checkpoint())
results = sess.run([boxes, scores, cropped_inputs],
feed_dict={inputs: image})
tf.logging.info("Successfully passed test.")
height = image.shape[1]
width = image.shape[2]
fig, ax = plt.subplots(1)
ax.imshow(image[0, :])
for i in range(results[0].shape[1]):
this_box = results[0][0, i, :]
box_y1 = this_box[0] * height
box_x1 = this_box[1] * width
box_y2 = this_box[2] * height
box_x2 = this_box[3] * width
rect = patches.Rectangle((box_x1, box_y1), (box_x2 - box_x1),
(box_y2 - box_y1),
linewidth=1,
edgecolor='r',
facecolor='none')
ax.add_patch(rect)
plt.savefig("images/image_boxes.png")
plt.clf()
for i in range(results[0].shape[1]):
fig, ax = plt.subplots(1)
ax.imshow(results[2][i, ...].astype(np.uint8))
plt.savefig("images/box{0}.png".format(i))
plt.clf()
def main(unused_argv):
image = load_image_from_path("images/image.jpg")[np.newaxis, ...]
vocab, pretrained_matrix = load_glove(vocab_size=100, embedding_size=50)
pos, pos_embeddings = get_parts_of_speech(), np.random.normal(
0, 0.1, [15, 50])
with tf.Graph().as_default():
inputs = tf.placeholder(tf.float32, shape=image.shape)
box_extractor = BoxExtractor(get_faster_rcnn_config(), top_k_boxes=16)
boxes, scores, cropped_inputs = box_extractor(inputs)
feature_extractor = FeatureExtractor()
mean_image_features = tf.reduce_mean(feature_extractor(inputs), [1, 2])
mean_object_features = tf.reshape(
tf.reduce_mean(feature_extractor(cropped_inputs), [1, 2]),
[1, 16, 2048])
image_captioner = PartOfSpeechImageCaptioner(UpDownCell(50), vocab,
pretrained_matrix,
UpDownCell(50),
UpDownCell(50), pos,
pos_embeddings)
pos_logits, pos_logits_ids, word_logits, word_logits_ids = image_captioner(
mean_image_features=mean_image_features,
mean_object_features=mean_object_features)
with tf.Session() as sess:
box_saver = tf.train.Saver(var_list=box_extractor.variables)
resnet_saver = tf.train.Saver(var_list=feature_extractor.variables)
box_saver.restore(sess, get_faster_rcnn_checkpoint())
resnet_saver.restore(sess, get_resnet_v2_101_checkpoint())
sess.run(tf.variables_initializer(image_captioner.variables))
results = sess.run(
[pos_logits, pos_logits_ids, word_logits, word_logits_ids],
feed_dict={inputs: image})
assert (results[2].shape[0] == 1 and results[2].shape[1] == 3
and results[2].shape[3] == 100)
tf.logging.info("Successfully passed test.")
({0:.2f} img/sec) image: {1:05d}
filename: {2}
attributes: {3}"""
tf.logging.set_verbosity(tf.logging.INFO)
tf.flags.DEFINE_string("file_pattern", "image.jpg", "")
FLAGS = tf.flags.FLAGS
if __name__ == "__main__":
attribute_map = get_visual_attributes()
with tf.Graph().as_default():
list_of_filenames = tf.gfile.Glob(FLAGS.file_pattern)
list_of_images = [
load_image_from_path(filename) for filename in list_of_filenames
]
box_extractor = BoxExtractor(get_faster_rcnn_config())
attribute_detector = AttributeDetector(1000)
feature_extractor = FeatureExtractor()
inputs = tf.placeholder(tf.float32, shape=[None, None, None, 3])
resized_inputs = tf.image.resize_images(inputs, [224, 224])
boxes, scores, cropped_inputs = box_extractor(inputs)
image_features = tf.reduce_mean(feature_extractor(resized_inputs),
[1, 2])
object_features = tf.reduce_mean(feature_extractor(cropped_inputs),
[1, 2])
batch_size = tf.shape(image_features)[0]
num_boxes = tf.shape(object_features)[0] // batch_size
depth = tf.shape(image_features)[1]