视频 API¶

注意

在 Colab 上试用或转到结尾下载完整的示例代码。

此示例说明了 torchvision 为视频提供的部分 API，以及有关如何构建数据集等的示例。

1. 简介：构建新的视频对象并检查属性¶

首先，我们选择一个视频来测试该对象。为了论证，我们使用 kinetics400 数据集中的一个视频。为了创建它，我们需要定义路径和要使用的流。

所选视频的统计信息

WUzgd7C1pWA.mp4
- 来源
  
  kinetics-400
- 视频
  
  H-264
  
  MPEG-4 AVC (第 10 部分) (avc1)
  
  fps：29.97
- 音频
  
  MPEG AAC 音频 (mp4a)
  
  采样率：48K Hz

import torch
import torchvision
from torchvision.datasets.utils import download_url
torchvision.set_video_backend("video_reader")

# Download the sample video
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/WUzgd7C1pWA.mp4?raw=true",
    ".",
    "WUzgd7C1pWA.mp4"
)
video_path = "./WUzgd7C1pWA.mp4"

Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/WUzgd7C1pWA.mp4 to ./WUzgd7C1pWA.mp4

7%
4%
1%
7%
4%
1%
8%
5%
2%
8%
5%
2%
9%
6%
3%
9%
6%
3%
0%
7%
4%
0%
7%
4%
1%
8%
5%
0%

流的定义方式类似于 torch 设备。我们将它们编码为字符串，形式为 stream_type:stream_id，其中 stream_type 是一个字符串，而 stream_id 是一个长整型。构造函数接受仅传递 stream_type，在这种情况下，流会自动发现。首先，让我们获取特定视频的元数据

stream = "video"
video = torchvision.io.VideoReader(video_path, stream)
video.get_metadata()

{'video': {'duration': [10.9109], 'fps': [29.97002997002997]}, 'audio': {'duration': [10.9], 'framerate': [48000.0]}, 'subtitles': {'duration': []}, 'cc': {'duration': []}}

这里我们可以看到视频有两个流 - 视频流和音频流。当前可用的流类型包括 [‘video’, ‘audio’]。每个描述符包含两部分：流类型（例如 ‘video’）和唯一的流 ID（由视频编码确定）。这样，如果视频容器包含多种相同类型的流，用户可以访问他们想要的流。如果只传递了流类型，解码器会自动检测到该类型的第一个流并将其返回。

让我们从视频流中读取所有帧。默认情况下，next(video_reader) 的返回值是一个包含以下字段的字典。

返回值字段为

data：包含一个 torch.tensor
pts：包含此特定帧的浮点时间戳

metadata = video.get_metadata()
video.set_current_stream("audio")

frames = []  # we are going to save the frames here.
ptss = []  # pts is a presentation timestamp in seconds (float) of each frame
for frame in video:
    frames.append(frame['data'])
    ptss.append(frame['pts'])

print("PTS for first five frames ", ptss[:5])
print("Total number of frames: ", len(frames))
approx_nf = metadata['audio']['duration'][0] * metadata['audio']['framerate'][0]
print("Approx total number of datapoints we can expect: ", approx_nf)
print("Read data size: ", frames[0].size(0) * len(frames))

PTS for first five frames  [0.0, 0.021332999999999998, 0.042667, 0.064, 0.08533299999999999]
Total number of frames:  511
Approx total number of datapoints we can expect:  523200.0
Read data size:  523264

但如果我们只想读取视频的特定时间段呢？这可以通过组合我们的 seek 函数以及每次调用 next 时以秒为单位返回的返回帧的表示时间戳来轻松完成。

鉴于我们的实现依赖于 python 迭代器，我们可以利用 itertools 来简化流程并使其更具 python 风格。

例如，如果我们想从第 2 秒读取 10 帧

import itertools
video.set_current_stream("video")

frames = []  # we are going to save the frames here.

# We seek into a second second of the video and use islice to get 10 frames since
for frame, pts in itertools.islice(video.seek(2), 10):
    frames.append(frame)

print("Total number of frames: ", len(frames))

Total number of frames:  10

或者如果我们想从第 2 秒到第 5 秒读取，我们会定位到视频的第 2 秒，然后利用 itertools takewhile 获取正确数量的帧

video.set_current_stream("video")
frames = []  # we are going to save the frames here.
video = video.seek(2)

for frame in itertools.takewhile(lambda x: x['pts'] <= 5, video):
    frames.append(frame['data'])

print("Total number of frames: ", len(frames))
approx_nf = (5 - 2) * video.get_metadata()['video']['fps'][0]
print("We can expect approx: ", approx_nf)
print("Tensor size: ", frames[0].size())

Total number of frames:  90
We can expect approx:  89.91008991008991
Tensor size:  torch.Size([3, 256, 340])

2. 构建示例 read_video 函数¶

我们可以利用以上方法来构建 read_video 函数，该函数遵循与现有 read_video 函数相同的 API。

def example_read_video(video_object, start=0, end=None, read_video=True, read_audio=True):
    if end is None:
        end = float("inf")
    if end < start:
        raise ValueError(
            "end time should be larger than start time, got "
            f"start time={start} and end time={end}"
        )

    video_frames = torch.empty(0)
    video_pts = []
    if read_video:
        video_object.set_current_stream("video")
        frames = []
        for frame in itertools.takewhile(lambda x: x['pts'] <= end, video_object.seek(start)):
            frames.append(frame['data'])
            video_pts.append(frame['pts'])
        if len(frames) > 0:
            video_frames = torch.stack(frames, 0)

    audio_frames = torch.empty(0)
    audio_pts = []
    if read_audio:
        video_object.set_current_stream("audio")
        frames = []
        for frame in itertools.takewhile(lambda x: x['pts'] <= end, video_object.seek(start)):
            frames.append(frame['data'])
            audio_pts.append(frame['pts'])
        if len(frames) > 0:
            audio_frames = torch.cat(frames, 0)

    return video_frames, audio_frames, (video_pts, audio_pts), video_object.get_metadata()


# Total number of frames should be 327 for video and 523264 datapoints for audio
vf, af, info, meta = example_read_video(video)
print(vf.size(), af.size())

torch.Size([327, 3, 256, 340]) torch.Size([523264, 1])

3. 构建示例随机采样数据集（可应用于 kinetics400 的训练数据集）¶

很酷，所以现在我们可以使用相同的原则来制作示例数据集。我们建议为此目的尝试使用可迭代数据集。这里，我们将构建一个示例数据集，该数据集读取随机选择的 10 帧视频。

制作示例数据集

import os
os.makedirs("./dataset", exist_ok=True)
os.makedirs("./dataset/1", exist_ok=True)
os.makedirs("./dataset/2", exist_ok=True)

下载视频

from torchvision.datasets.utils import download_url
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/WUzgd7C1pWA.mp4?raw=true",
    "./dataset/1", "WUzgd7C1pWA.mp4"
)
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/RATRACE_wave_f_nm_np1_fr_goo_37.avi?raw=true",
    "./dataset/1",
    "RATRACE_wave_f_nm_np1_fr_goo_37.avi"
)
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/SOX5yA1l24A.mp4?raw=true",
    "./dataset/2",
    "SOX5yA1l24A.mp4"
)
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/v_SoccerJuggling_g23_c01.avi?raw=true",
    "./dataset/2",
    "v_SoccerJuggling_g23_c01.avi"
)
download_url(
    "https://github.com/pytorch/vision/blob/main/test/assets/videos/v_SoccerJuggling_g24_c01.avi?raw=true",
    "./dataset/2",
    "v_SoccerJuggling_g24_c01.avi"
)

Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/WUzgd7C1pWA.mp4 to ./dataset/1/WUzgd7C1pWA.mp4

7%
4%
1%
7%
4%
1%
8%
5%
2%
8%
5%
2%
9%
6%
3%
9%
6%
3%
0%
7%
4%
0%
7%
4%
1%
8%
5%
0%
Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/RATRACE_wave_f_nm_np1_fr_goo_37.avi to ./dataset/1/RATRACE_wave_f_nm_np1_fr_goo_37.avi

4%
9%
3%
7%
1%
6%
0%
4%
0%
Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/SOX5yA1l24A.mp4 to ./dataset/2/SOX5yA1l24A.mp4

8%
7%
5%
4%
2%
1%
9%
8%
6%
5%
3%
2%
0%
9%
7%
6%
4%
0%
Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/v_SoccerJuggling_g23_c01.avi to ./dataset/2/v_SoccerJuggling_g23_c01.avi

4%
9%
3%
8%
2%
7%
1%
6%
0%
5%
9%
3%
8%
2%
7%
0%
Downloading https://raw.githubusercontent.com/pytorch/vision/refs/heads/main/test/assets/videos/v_SoccerJuggling_g24_c01.avi to ./dataset/2/v_SoccerJuggling_g24_c01.avi

3%
5%
8%
0%
3%
6%
8%
1%
3%
6%
9%
1%
4%
6%
9%
2%
4%
7%
9%
0%

家政和实用程序

import os
import random

from torchvision.datasets.folder import make_dataset
from torchvision import transforms as t


def _find_classes(dir):
    classes = [d.name for d in os.scandir(dir) if d.is_dir()]
    classes.sort()
    class_to_idx = {cls_name: i for i, cls_name in enumerate(classes)}
    return classes, class_to_idx


def get_samples(root, extensions=(".mp4", ".avi")):
    _, class_to_idx = _find_classes(root)
    return make_dataset(root, class_to_idx, extensions=extensions)

我们将定义数据集和一些基本参数。我们假设 FolderDataset 的结构，并添加以下参数

clip_len：以帧为单位的剪辑长度
frame_transform：对每帧单独进行的转换
video_transform：对视频序列进行的转换

注意

实际上，我们将时期大小添加为使用 IterableDataset() 类允许我们根据需要自然地对每个视频的剪辑或图像进行过采样。

class RandomDataset(torch.utils.data.IterableDataset):
    def __init__(self, root, epoch_size=None, frame_transform=None, video_transform=None, clip_len=16):
        super(RandomDataset).__init__()

        self.samples = get_samples(root)

        # Allow for temporal jittering
        if epoch_size is None:
            epoch_size = len(self.samples)
        self.epoch_size = epoch_size

        self.clip_len = clip_len
        self.frame_transform = frame_transform
        self.video_transform = video_transform

    def __iter__(self):
        for i in range(self.epoch_size):
            # Get random sample
            path, target = random.choice(self.samples)
            # Get video object
            vid = torchvision.io.VideoReader(path, "video")
            metadata = vid.get_metadata()
            video_frames = []  # video frame buffer

            # Seek and return frames
            max_seek = metadata["video"]['duration'][0] - (self.clip_len / metadata["video"]['fps'][0])
            start = random.uniform(0., max_seek)
            for frame in itertools.islice(vid.seek(start), self.clip_len):
                video_frames.append(self.frame_transform(frame['data']))
                current_pts = frame['pts']
            # Stack it into a tensor
            video = torch.stack(video_frames, 0)
            if self.video_transform:
                video = self.video_transform(video)
            output = {
                'path': path,
                'video': video,
                'target': target,
                'start': start,
                'end': current_pts}
            yield output

给定文件夹结构中视频的路径，即

数据集
- 类别 1
  
  文件 0
  
  文件 1
  
  …
- 类别 2
  
  文件 0
  
  文件 1
  
  …
- …

我们可以生成一个 dataloader 并测试数据集。

transforms = [t.Resize((112, 112))]
frame_transform = t.Compose(transforms)

dataset = RandomDataset("./dataset", epoch_size=None, frame_transform=frame_transform)

from torch.utils.data import DataLoader
loader = DataLoader(dataset, batch_size=12)
data = {"video": [], 'start': [], 'end': [], 'tensorsize': []}
for batch in loader:
    for i in range(len(batch['path'])):
        data['video'].append(batch['path'][i])
        data['start'].append(batch['start'][i].item())
        data['end'].append(batch['end'][i].item())
        data['tensorsize'].append(batch['video'][i].size())
print(data)

{'video': ['./dataset/1/RATRACE_wave_f_nm_np1_fr_goo_37.avi', './dataset/1/RATRACE_wave_f_nm_np1_fr_goo_37.avi', './dataset/1/WUzgd7C1pWA.mp4', './dataset/2/SOX5yA1l24A.mp4', './dataset/1/RATRACE_wave_f_nm_np1_fr_goo_37.avi'], 'start': [1.203051270746008, 0.5760754748483161, 10.23800984898201, 9.709010060342672, 1.256045985643026], 'end': [1.733333, 1.0999999999999999, 10.744067, 10.2102, 1.766667], 'tensorsize': [torch.Size([16, 3, 112, 112]), torch.Size([16, 3, 112, 112]), torch.Size([16, 3, 112, 112]), torch.Size([16, 3, 112, 112]), torch.Size([16, 3, 112, 112])]}

4. 数据可视化¶

可视化视频的示例

import matplotlib.pyplot as plt

plt.figure(figsize=(12, 12))
for i in range(16):
    plt.subplot(4, 4, i + 1)
    plt.imshow(batch["video"][0, i, ...].permute(1, 2, 0))
    plt.axis("off")

清理视频和数据集

import os
import shutil
os.remove("./WUzgd7C1pWA.mp4")
shutil.rmtree("./dataset")

脚本的总运行时间：（0 分钟 4.939 秒）

由 Sphinx-Gallery 生成的图库