def display_image_id(self, image_id, json_data):
path = str('./Dataset_COCO_5k_images')
img_id_annotation = 0
count = 0
for i in json_data['annotations']:
if (json_data['annotations'][count]['image_id'] == image_id):
img_id_annotation = count
break
count += 1
image = None # Image-object
image_caption = None
for filename in os.listdir(path):
if (int(filename[1:-4]) == json_data['annotations']
[img_id_annotation]['image_id']):
image = Image.open(
os.path.join('Dataset_COCO_5k_images', filename))
image_caption = json_data['annotations'][img_id_annotation][
'caption']
print('Image-Filename: {}'.format(filename))
print('Image-shape: {}'.format(image.size))
print('Image-captioning: {}'.format(image_caption))
plt.imshow(image)
plt.show()
image.close()
break
def generate_samples_end_to_end(filename, patch_size, seed, batch_size=100):
prng = random.Random()
prng.seed(seed)
batch = ([], [])
with tarfile.open(filename, 'r:*') as tar:
entries = [x.name for x in tar.getmembers() if x.isreg()]
while True:
random.shuffle(entries)
for thing in entries:
with tar.extractfile(thing) as jpeg:
image = Image.open(jpeg)
image.load()
try:
pix = np.array(image)
dynamic_range_factor = 0.99
offset = (1 - dynamic_range_factor) / 2
if 3 == len(
pix.shape
) and 3 == pix.shape[2] and pix.shape[0] > patch_size[
0] and pix.shape[1] > patch_size[1]:
for i in range(10):
offset_row = math.floor(
prng.uniform(0,
pix.shape[0] - patch_size[0]))
offset_col = math.floor(
prng.uniform(0,
pix.shape[1] - patch_size[1]))
chip = dynamic_range_factor * pix[offset_row:(
offset_row + patch_size[0]), offset_col:(
offset_col + patch_size[1]), :].copy(
).astype(np.float32) / 255 + offset
if (patch_size[0], patch_size[1],
3) == chip.shape:
batch[0].append(chip)
batch[1].append(chip)
if len(batch[0]) == batch_size:
yield np.stack(batch[0],
axis=0), np.stack(
batch[1], axis=0)
batch[0].clear()
batch[1].clear()
gc.collect()
del pix
finally:
image.close()
del image
def COCO_images_segmentation(self, num_img=100):
image_array_segmentation = []
path = str('./Dataset_COCO_5k_images_segmentation')
count = 0
for filename in tqdm(os.listdir(path)):
if (count >= num_img):
break
image = Image.open(
os.path.join('Dataset_COCO_5k_images_segmentation', filename))
image_array_segmentation.append(filename)
image.close()
count += 1
return image_array_segmentation
int(100 * scores[j]))
print('cropping box {0} ({1}, {2}, {3}, {4})'.format(
display_str, xmin_px, ymin_px, xmax_px, ymax_px))
cropped_img = image.crop(
(xmin_px, ymin_px, xmax_px, ymax_px))
with tf.gfile.Open(
os.path.join(
PATH_TO_TEST_IMAGES_CROP_RES_DIR,
'{0}_{1}_{2}.jpg'.format(
image_path_filename, display_str, j)),
'w') as fid:
cropped_img.save(fid, 'JPEG')
features = get_feature(cropped_img)
w.writerow({
"filename": image_path,
"object_class": class_name,
"score": scores[j],
"ymin": ymin,
"xmin": xmin,
"ymax": ymax,
"xmax": xmax,
"features": ','.join(features)
})
cropped_img.close()
image.close()