O *Dataset* de Detecção de Objetos
==================================
Não existem pequenos conjuntos de dados, como MNIST ou Fashion-MNIST, no
campo de detecção de objetos. Para testar os modelos rapidamente, vamos
montar um pequeno conjunto de dados. Primeiro, geramos 1000 imagens de
banana de diferentes ângulos e tamanhos usando bananas grátis de nosso
escritório. Em seguida, coletamos uma série de imagens de fundo e
colocamos uma imagem de banana em uma posição aleatória em cada imagem.
Baixando Dataset
----------------
O conjunto de dados de detecção de banana com todas as imagens e
arquivos de rótulo csv pode ser baixado diretamente da Internet.
.. raw:: html
.. raw:: html
.. code:: python
%matplotlib inline
import os
import pandas as pd
from mxnet import gluon, image, np, npx
from d2l import mxnet as d2l
npx.set_np()
#@save
d2l.DATA_HUB['banana-detection'] = (d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. code:: python
%matplotlib inline
import os
import pandas as pd
import torch
import torchvision
from d2l import torch as d2l
#@save
d2l.DATA_HUB['banana-detection'] = (d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
#@save
def read_data_bananas(is_train=True):
"""Read the bananas dataset images and labels."""
data_dir = d2l.download_extract('banana-detection')
csv_fname = os.path.join(data_dir, 'bananas_train' if is_train
else 'bananas_val', 'label.csv')
csv_data = pd.read_csv(csv_fname)
csv_data = csv_data.set_index('img_name')
images, targets = [], []
for img_name, target in csv_data.iterrows():
images.append(image.imread(
os.path.join(data_dir, 'bananas_train' if is_train else
'bananas_val', 'images', f'{img_name}')))
# Since all images have same object class i.e. category '0',
# the `label` column corresponds to the only object i.e. banana
# The target is as follows : (`label`, `xmin`, `ymin`, `xmax`, `ymax`)
targets.append(list(target))
return images, np.expand_dims(np.array(targets), 1) / 256
#@save
class BananasDataset(gluon.data.Dataset):
def __init__(self, is_train):
self.features, self.labels = read_data_bananas(is_train)
print('read ' + str(len(self.features)) + (f' training examples' if
is_train else f' validation examples'))
def __getitem__(self, idx):
return (self.features[idx].astype('float32').transpose(2, 0, 1),
self.labels[idx])
def __len__(self):
return len(self.features)
#@save
def load_data_bananas(batch_size):
"""Load the bananas dataset."""
train_iter = gluon.data.DataLoader(BananasDataset(is_train=True),
batch_size, shuffle=True)
val_iter = gluon.data.DataLoader(BananasDataset(is_train=False),
batch_size)
return (train_iter, val_iter)
.. raw:: html
.. raw:: html
.. code:: python
#@save
def read_data_bananas(is_train=True):
"""Read the bananas dataset images and labels."""
data_dir = d2l.download_extract('banana-detection')
csv_fname = os.path.join(data_dir, 'bananas_train' if is_train
else 'bananas_val', 'label.csv')
csv_data = pd.read_csv(csv_fname)
csv_data = csv_data.set_index('img_name')
images, targets = [], []
for img_name, target in csv_data.iterrows():
images.append(torchvision.io.read_image(
os.path.join(data_dir, 'bananas_train' if is_train else
'bananas_val', 'images', f'{img_name}')))
# Since all images have same object class i.e. category '0',
# the `label` column corresponds to the only object i.e. banana
# The target is as follows : (`label`, `xmin`, `ymin`, `xmax`, `ymax`)
targets.append(list(target))
return images, torch.tensor(targets).unsqueeze(1) / 256
#@save
class BananasDataset(torch.utils.data.Dataset):
def __init__(self, is_train):
self.features, self.labels = read_data_bananas(is_train)
print('read ' + str(len(self.features)) + (f' training examples' if
is_train else f' validation examples'))
def __getitem__(self, idx):
return (self.features[idx].float(), self.labels[idx])
def __len__(self):
return len(self.features)
#@save
def load_data_bananas(batch_size):
"""Load the bananas dataset."""
train_iter = torch.utils.data.DataLoader(BananasDataset(is_train=True),
batch_size, shuffle=True)
val_iter = torch.utils.data.DataLoader(BananasDataset(is_train=False),
batch_size)
return (train_iter, val_iter)
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
batch_size, edge_size = 32, 256
train_iter, _ = load_data_bananas(batch_size)
batch = next(iter(train_iter))
batch[0].shape, batch[1].shape
.. parsed-literal::
:class: output
Downloading ../data/banana-detection.zip from http://d2l-data.s3-accelerate.amazonaws.com/banana-detection.zip...
read 1000 training examples
read 100 validation examples
.. parsed-literal::
:class: output
((32, 3, 256, 256), (32, 1, 5))
.. raw:: html
.. raw:: html
.. code:: python
batch_size, edge_size = 32, 256
train_iter, _ = load_data_bananas(batch_size)
batch = next(iter(train_iter))
batch[0].shape, batch[1].shape
.. parsed-literal::
:class: output
Downloading ../data/banana-detection.zip from http://d2l-data.s3-accelerate.amazonaws.com/banana-detection.zip...
read 1000 training examples
read 100 validation examples
.. parsed-literal::
:class: output
(torch.Size([32, 3, 256, 256]), torch.Size([32, 1, 5]))
.. raw:: html
.. raw:: html
.. raw:: html
.. code:: python
imgs = (batch[0][0:10].transpose(0, 2, 3, 1)) / 255
axes = d2l.show_images(imgs, 2, 5, scale=2)
for ax, label in zip(axes, batch[1][0:10]):
d2l.show_bboxes(ax, [label[0][1:5] * edge_size], colors=['w'])
.. figure:: output_object-detection-dataset_641ef0_30_0.png
.. raw:: html
.. raw:: html
.. code:: python
imgs = (batch[0][0:10].permute(0, 2, 3, 1)) / 255
axes = d2l.show_images(imgs, 2, 5, scale=2)
for ax, label in zip(axes, batch[1][0:10]):
d2l.show_bboxes(ax, [label[0][1:5] * edge_size], colors=['w'])
.. figure:: output_object-detection-dataset_641ef0_33_0.png
.. raw:: html
.. raw:: html
.. raw:: html
`Discussões `__
.. raw:: html
.. raw:: html
`Discussões `__
.. raw:: html
.. raw:: html