CocoDetection in PyTorch (1)

Buy Me a Coffee☕

*My post explains MS COCO.

CocoDetection() can use MS COCO dataset as shown below:

*Memos:

• The 1st argument is root(Required-Type:str or pathlib.Path): *Memos:
  • It's the path to the images.
  • An absolute or relative path is possible.
• The 2nd argument is annFile(Required-Type:str or pathlib.Path): *Memos:
  • It's the path to the annotations.
  • An absolute or relative path is possible.
• The 3rd argument is transform(Optional-Default:None-Type:callable).
• The 4th argument is target_transform(Optional-Default:None-Type:callable).
• The 5th argument is transforms(Optional-Default:None-Type:callable).

from torchvision.datasets import CocoDetection

cap_train2014_data = CocoDetection(
    root="data/coco/imgs/train2014",
    annFile="data/coco/anns/trainval2014/captions_train2014.json"
)

cap_train2014_data = CocoDetection(
    root="data/coco/imgs/train2014",
    annFile="data/coco/anns/trainval2014/captions_train2014.json",
    transform=None,
    target_transform=None,
    transforms=None
)

ins_train2014_data = CocoDetection(
    root="data/coco/imgs/train2014",
    annFile="data/coco/anns/trainval2014/instances_train2014.json"
)

pk_train2014_data = CocoDetection(
    root="data/coco/imgs/train2014",
    annFile="data/coco/anns/trainval2014/person_keypoints_train2014.json"
)

len(cap_train2014_data), len(ins_train2014_data), len(pk_train2014_data)
# (82783, 82783, 82783)

cap_val2014_data = CocoDetection(
    root="data/coco/imgs/val2014",
    annFile="data/coco/anns/trainval2014/captions_val2014.json"
)

ins_val2014_data = CocoDetection(
    root="data/coco/imgs/val2014",
    annFile="data/coco/anns/trainval2014/instances_val2014.json"
)

pk_val2014_data = CocoDetection(
    root="data/coco/imgs/val2014",
    annFile="data/coco/anns/trainval2014/person_keypoints_val2014.json"
)

len(cap_val2014_data), len(ins_val2014_data), len(pk_val2014_data)
# (40504, 40504, 40504)

test2014_data = CocoDetection(
    root="data/coco/imgs/test2014",
    annFile="data/coco/anns/test2014/test2014.json"
)

test2015_data = CocoDetection(
    root="data/coco/imgs/test2015",
    annFile="data/coco/anns/test2015/test2015.json"
)

testdev2015_data = CocoDetection(
    root="data/coco/imgs/test2015",
    annFile="data/coco/anns/test2015/test-dev2015.json"
)

len(test2014_data), len(test2015_data), len(testdev2015_data)
# (40775, 81434, 20288)

cap_train2014_data
# Dataset CocoDetection
#     Number of datapoints: 82783
#     Root location: data/coco/imgs/train2014

cap_train2014_data.root
# 'data/coco/imgs/train2014'

print(cap_train2014_data.transform)
# None

print(cap_train2014_data.target_transform)
# None

print(cap_train2014_data.transforms)
# None

cap_train2014_data[0]
# (<pil.image.image image mode="RGB" size="640x480">,
#  [{'image_id': 9, 'id': 661611,
#    'caption': 'Closeup of bins of food that include broccoli and bread.'},
#   {'image_id': 9, 'id': 661977,
#    'caption': 'A meal is presented in brightly colored plastic trays.'},
#   {'image_id': 9, 'id': 663627,
#    'caption': 'there are containers filled with different kinds of foods'},
#   {'image_id': 9, 'id': 666765,
#    'caption': 'Colorful dishes holding meat, vegetables, fruit, and bread.'},
#   {'image_id': 9, 'id': 667602,
#    'caption': 'A bunch of trays that have different food.'}])　

cap_train2014_data[1]
# (<pil.image.image image mode="RGB" size="640x426">,
#  [{'image_id': 25, 'id': 122312,
#    'caption': 'A giraffe eating food from the top of the tree.'},
#   {'image_id': 25, 'id': 127076,
#    'caption': 'A giraffe standing up nearby a tree '},
#   {'image_id': 25, 'id': 127238,
#    'caption': 'A giraffe mother with its baby in the forest.'},
#   {'image_id': 25, 'id': 133058,
#    'caption': 'Two giraffes standing in a tree filled area.'},
#   {'image_id': 25, 'id': 133676,
#    'caption': 'A giraffe standing next to a forest filled with trees.'}])

cap_train2014_data[2]
# (<pil.image.image image mode="RGB" size="640x428">,
#  [{'image_id': 30, 'id': 695774,
#    'caption': 'A flower vase is sitting on a porch stand.'},
#   {'image_id': 30, 'id': 696557,
#    'caption': 'White vase with different colored flowers sitting inside of it. '},
#   {'image_id': 30, 'id': 699041,
#    'caption': 'a white vase with many flowers on a stage'},
#   {'image_id': 30, 'id': 701216,
#    'caption': 'A white vase filled with different colored flowers.'},
#   {'image_id': 30, 'id': 702428,
#    'caption': 'A vase with red and white flowers outside on a sunny day.'}])

ins_train2014_data[0]
# (<pil.image.image image mode="RGB" size="640x480">,
#  [{'segmentation': [[500.49, 473.53, 599.73, ..., 20.49, 473.53]],
#    'area': 120057.13925, 'iscrowd': 0, 'image_id': 9,
#    'bbox': [1.08, 187.69, 611.59, 285.84], 'category_id': 51,
#    'id': 1038967},
#   {'segmentation': ..., 'category_id': 51, 'id': 1039564},
#   ...,
#   {'segmentation': ..., 'category_id': 55, 'id': 1914001}])

ins_train2014_data[1]
# (<pil.image.image image mode="RGB" size="640x426">,
#  [{'segmentation': [[437.52, 353.33, 437.87, ..., 437.87, 357.19]],
#    'area': 19686.597949999996, 'iscrowd': 0, 'image_id': 25,
#    'bbox': [385.53, 60.03, 214.97, 297.16], 'category_id': 25,
#    'id': 598548},
#  {'segmentation': [[99.26, 405.72, 133.57, ..., 97.77, 406.46]],
#   'area': 2785.8475500000004, 'iscrowd': 0, 'image_id': 25,
#   'bbox': [53.01, 356.49, 132.03, 55.19], 'category_id': 25,
#   'id': 599491}])

ins_train2014_data[2]
# (<pil.image.image image mode="RGB" size="640x428">,
#  [{'segmentation': [[267.38, 330.14, 281.81, ..., 269.3, 329.18]],
#    'area': 47675.66289999999, 'iscrowd': 0, 'image_id': 30,
#    'bbox': [204.86, 31.02, 254.88, 324.12], 'category_id': 64,
#    'id': 291613},
#   {'segmentation': [[394.34, 155.81, 403.96, ..., 393.38, 157.73]],
#    'area': 16202.798250000003, 'iscrowd': 0, 'image_id': 30,
#    'bbox': [237.56, 155.81, 166.4, 195.25], 'category_id': 86,
#    'id': 1155486}])

pk_train2014_data[0]
# (<pil.image.image image mode="RGB" size="640x480">, [])

pk_train2014_data[1]
# (<pil.image.image image mode="RGB" size="640x426">, [])

pk_train2014_data[2]
# (<pil.image.image image mode="RGB" size="640x428">, [])

cap_val2014_data[0]
# (<pil.image.image image mode="RGB" size="640x478">,
#  [{'image_id': 42, 'id': 641613,
#    'caption': 'This wire metal rack holds several pairs of shoes and sandals'},
#   {'image_id': 42, 'id': 645309,
#    'caption': 'A dog sleeping on a show rack in the shoes.'},
#   {'image_id': 42, 'id': 650217,
#    'caption': 'Various slides and other footwear rest in a metal basket outdoors.'},
#   {'image_id': 42,
#    'id': 650868,
#    'caption': 'A small dog is curled up on top of the shoes'},
#   {'image_id': 42,
#    'id': 652383,
#    'caption': 'a shoe rack with some shoes and a dog sleeping on them'}])

cap_val2014_data[1]
# (<pil.image.image image mode="RGB" size="565x640">,
#  [{'image_id': 73, 'id': 593422,
#    'caption': 'A motorcycle parked in a parking space next to another motorcycle.'},
#   {'image_id': 73, 'id': 746071,
#    'caption': 'An old motorcycle parked beside other motorcycles with a brown leather seat.'},
#   {'image_id': 73, 'id': 746170,
#    'caption': 'Motorcycle parked in the parking lot of asphalt.'},
#   {'image_id': 73, 'id': 746914,
#    'caption': 'A close up view of a motorized bicycle, sitting in a rack. '},
#   {'image_id': 73, 'id': 748185,
#    'caption': 'The back tire of an old style motorcycle is resting in a metal stand. '}])

cap_val2014_data[2]
# (<pil.image.image image mode="RGB" size="640x426">,
#  [{'image_id': 74, 'id': 145996,
#    'caption': 'A picture of a dog laying on the ground.'},
#   {'image_id': 74, 'id': 146710,
#    'caption': 'Dog snoozing by a bike on the edge of a cobblestone street'},
#   {'image_id': 74, 'id': 149398,
#    'caption': 'The white dog lays next to the bicycle on the sidewalk.'},
#   {'image_id': 74, 'id': 149638,
#    'caption': 'a white dog is sleeping on a street and a bicycle'},
#   {'image_id': 74, 'id': 150181,
#    'caption': 'A puppy rests on the street next to a bicycle.'}])

ins_val2014_data[0]
# (<pil.image.image image mode="RGB" size="640x478">,
#  [{'segmentation': [[382.48, 268.63, 330.24, ..., 394.09, 264.76]],
#    'area': 53481.5118, 'iscrowd': 0, 'image_id': 42,
#    'bbox': [214.15, 41.29, 348.26, 243.78], 'category_id': 18,
#    'id': 1817255}])

ins_val2014_data[1]
# (<pil.image.image image mode="RGB" size="565x640">,
#  [{'segmentation': [[134.36, 145.55, 117.02, ..., 138.69, 141.22]],
#    'area': 172022.43864999997, 'iscrowd': 0, 'image_id': 73,
#    'bbox': [13.0, 22.75, 535.98, 609.67], 'category_id': 4,
#    'id': 246920},
#   {'segmentation': [[202.28, 4.97, 210.57, 26.53, ..., 192.33, 3.32]],
#    'area': 52666.3402, 'iscrowd': 0, 'image_id': 73,
#    'bbox': [1.66, 3.32, 268.6, 271.91], 'category_id': 4,
#    'id': 2047387}])

ins_val2014_data[2]
# (<pil.image.image image mode="RGB" size="640x426">,
#  [{'segmentation': [[321.02, 321.0, 314.25, ..., 320.57, 322.86]],
#    'area': 18234.62355, 'iscrowd': 0, 'image_id': 74,
#    'bbox': [61.87, 276.25, 296.42, 103.18], 'category_id': 18,
#    'id': 1774},
#   {'segmentation': ..., 'category_id': 2, 'id': 128367},
#   ...
#   {'segmentation': ..., 'category_id': 1, 'id': 1751664}])

pk_val2014_data[0]
# (<pil.image.image image mode="RGB" size="640x478">, [])

pk_val2014_data[1]
# (<pil.image.image image mode="RGB" size="565x640">, [])

pk_val2014_data[2]
# (<pil.image.image image mode="RGB" size="640x426">,
#  [{'segmentation': [[301.32, 93.96, 305.72, ..., 299.67, 94.51]],
#    'num_keypoints': 0, 'area': 638.7158, 'iscrowd': 0,
#    'keypoints': [0, 0, 0, 0, ..., 0, 0], 'image_id': 74,
#    'bbox': [295.55, 93.96, 18.42, 58.83], 'category_id': 1,
#    'id': 195946},
#   {'segmentation': ..., 'category_id': 1, 'id': 253933},
#   ...
#   {'segmentation': ..., 'category_id': 1, 'id': 1751664}])

test2014_data[0]
# (<pil.image.image image mode="RGB" size="640x480">, [])

test2014_data[1]
# (<pil.image.image image mode="RGB" size="480x640">, [])

test2014_data[2]
# (<pil.image.image image mode="RGB" size="480x640">, [])

test2015_data[0]
# (<pil.image.image image mode="RGB" size="640x480">, [])

test2015_data[1]
# (<pil.image.image image mode="RGB" size="480x640">, [])

test2015_data[2]
# (<pil.image.image image mode="RGB" size="480x640">, [])

testdev2015_data[0]
# (<pil.image.image image mode="RGB" size="640x480">, [])

testdev2015_data[1]
# (<pil.image.image image mode="RGB" size="480x640">, [])

testdev2015_data[2]
# (<pil.image.image image mode="RGB" size="640x427">, [])

import matplotlib.pyplot as plt
from matplotlib.patches import Polygon, Rectangle
import torch

def show_images(data, main_title=None):
    file = data.root.split('/')[-1]
    if data[0][1] and "caption" in data[0][1][0]:
        if file == "train2014":
            plt.figure(figsize=(14, 5))
            plt.suptitle(t=main_title, y=0.9, fontsize=14)
            x_axis = 0.02
            x_axis_incr = 0.325
            fs = 10.5
        elif file == "val2014":
            plt.figure(figsize=(14, 6.5))
            plt.suptitle(t=main_title, y=0.94, fontsize=14)
            x_axis = 0.01
            x_axis_incr = 0.32
            fs = 9.4
        for i, (im, ann) in zip(range(1, 4), data):
            plt.subplot(1, 3, i)
            plt.imshow(X=im)
            plt.title(label=ann[0]["image_id"])
            y_axis = 0.0
            for j in range(0, 5):
                plt.figtext(x=x_axis, y=y_axis, fontsize=fs,
                            s=f'{ann[j]["id"]}:\n{ann[j]["caption"]}')
                if file == "train2014":
                    y_axis -= 0.1
                elif file == "val2014":
                    y_axis -= 0.07
            x_axis += x_axis_incr
            if i == 2 and file == "val2014":
                x_axis += 0.06
        plt.tight_layout()
        plt.show()
    elif data[0][1] and "segmentation" in data[0][1][0]:
        if file == "train2014":
            fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(14, 4))
        elif file == "val2014":
            fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(14, 5))
        fig.suptitle(t=main_title, y=1.0, fontsize=14)
        for (im, anns), axis in zip(data, axes.ravel()):
            for ann in anns:
                for seg in ann['segmentation']:
                    seg_tsors = torch.tensor(seg).split(2)
                    seg_lists = [seg_tsor.tolist() for seg_tsor in seg_tsors]
                    poly = Polygon(xy=seg_lists,
                                   facecolor="lightgreen", alpha=0.7)
                    axis.add_patch(p=poly)
                    px = []
                    py = []
                    for j, v in enumerate(seg):
                        if j%2 == 0:
                            px.append(v)
                        else:
                            py.append(v)
                    axis.plot(px, py, color='yellow')
                x, y, w, h = ann['bbox']
                rect = Rectangle(xy=(x, y), width=w, height=h,
                                 linewidth=3, edgecolor='r',
                                 facecolor='none', zorder=2)
                axis.add_patch(p=rect)
            axis.imshow(X=im)
            axis.set_title(label=anns[0]["image_id"])
        fig.tight_layout()
        plt.show()
    elif not data[0][1]:
        if file == "train2014":
            plt.figure(figsize=(14, 5))
            plt.suptitle(t=main_title, y=0.9, fontsize=14)
        elif file == "val2014":
            plt.figure(figsize=(14, 5))
            plt.suptitle(t=main_title, y=1.05, fontsize=14)
        elif file == "test2014" or "test2015":
            plt.figure(figsize=(14, 8))
            plt.suptitle(t=main_title, y=0.9, fontsize=14)
        for i, (im, _) in zip(range(1, 4), data):
            plt.subplot(1, 3, i)
            plt.imshow(X=im)
        plt.tight_layout()
        plt.show()

show_images(data=cap_train2014_data, main_title="cap_train2014_data")
show_images(data=ins_train2014_data, main_title="ins_train2014_data")
show_images(data=pk_train2014_data, main_title="pk_train2014_data")

show_images(data=cap_val2014_data, main_title="cap_val2014_data")
show_images(data=ins_val2014_data, main_title="ins_val2014_data")
show_images(data=pk_val2014_data, main_title="pk_val2014_data")

show_images(data=test2014_data, main_title="test2014_data")
show_images(data=test2015_data, main_title="test2015_data")
show_images(data=testdev2015_data, main_title="testdev2015_data")
</pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image></pil.image.image>

The above is the detailed content of CocoDetection in PyTorch (1). For more information, please follow other related articles on the PHP Chinese website!