.. _sec-object-detection-dataset:
物体検出データセット
====================
物体検出の分野には、MNIST や Fashion-MNIST
のような小規模データセットはない。 物体検出モデルを手早く示すために、
私たちは小さなデータセットを収集してラベル付けした。
まず、オフィスにあった自由に使えるバナナの写真を撮り、
回転やサイズの異なる 1000 枚のバナナ画像を生成した。
次に、それぞれのバナナ画像を
いくつかの背景画像上のランダムな位置に配置した。
最後に、それらの画像中のバナナに対してバウンディングボックスをラベル付けした。
データセットのダウンロード
--------------------------
画像ファイルと csv
ラベルファイルをすべて含むバナナ検出データセットは、インターネットから直接ダウンロードできる。
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
%matplotlib inline
from d2l import torch as d2l
import torch
import torchvision
import os
import pandas as pd
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
%matplotlib inline
from d2l import mxnet as d2l
from mxnet import gluon, image, np, npx
import os
import pandas as pd
npx.set_np()
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
d2l.DATA_HUB['banana-detection'] = (
d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
d2l.DATA_HUB['banana-detection'] = (
d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
d2l.DATA_HUB['banana-detection'] = (
d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
d2l.DATA_HUB['banana-detection'] = (
d2l.DATA_URL + 'banana-detection.zip',
'5de26c8fce5ccdea9f91267273464dc968d20d72')
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. 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
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. 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
.. raw:: html
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
def read_data_bananas(is_train=True):
"""Read the banana detection 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}')))
# Here `target` contains (class, upper-left x, upper-left y,
# lower-right x, lower-right y), where all the images have the same
# banana class (index 0)
targets.append(list(target))
return images, torch.tensor(targets).unsqueeze(1) / 256
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
def read_data_bananas(is_train=True):
"""Read the banana detection 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}')))
# Here `target` contains (class, upper-left x, upper-left y,
# lower-right x, lower-right y), where all the images have the same
# banana class (index 0)
targets.append(list(target))
return images, np.expand_dims(np.array(targets), 1) / 256
.. raw:: html
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
class BananasDataset(torch.utils.data.Dataset):
"""A customized dataset to load the banana detection 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)
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
class BananasDataset(gluon.data.Dataset):
"""A customized dataset to load the banana detection 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)
.. raw:: html
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
#@save
def load_data_bananas(batch_size):
"""Load the banana detection 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:: latex
\diilbookstyleinputcell
.. code:: python
#@save
def load_data_bananas(batch_size):
"""Load the banana detection 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
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
imgs = (batch[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][:10]):
d2l.show_bboxes(ax, [label[0][1:5] * edge_size], colors=['w'])
.. figure:: output_object-detection-dataset_42a473_61_0.png
.. raw:: html
.. raw:: html
.. raw:: latex
\diilbookstyleinputcell
.. code:: python
imgs = (batch[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][:10]):
d2l.show_bboxes(ax, [label[0][1:5] * edge_size], colors=['w'])
.. figure:: output_object-detection-dataset_42a473_64_0.png
.. raw:: html
.. raw:: html