动作识别

动作识别可识别与视频剪辑不同的动作，例如步行或跳舞。整个视频持续时间内不一定每个动作都执行。

使用 AutoML 模型

准备工作

如需了解创建 AutoML 模型的背景信息，请查看 Vertex AI 新手指南。如需了解如何创建 AutoML 模型，请参阅 Vertex AI 文档“开发和使用机器学习模型”中的视频数据。

使用您的 AutoML 模型

以下代码示例演示了如何使用 AutoML 模型，通过流式客户端库执行操作识别。

Python

要向 Video Intelligence 进行身份验证，请设置应用默认凭据。如需了解详情，请参阅为本地开发环境设置身份验证。 

import io

from google.cloud import videointelligence_v1p3beta1 as videointelligence

# path = 'path_to_file'
# project_id = 'project_id'
# model_id = 'automl_action_recognition_model_id'

client = videointelligence.StreamingVideoIntelligenceServiceClient()

model_path = "projects/{}/locations/us-central1/models/{}".format(
    project_id, model_id
)

automl_config = videointelligence.StreamingAutomlActionRecognitionConfig(
    model_name=model_path
)

video_config = videointelligence.StreamingVideoConfig(
    feature=videointelligence.StreamingFeature.STREAMING_AUTOML_ACTION_RECOGNITION,
    automl_action_recognition_config=automl_config,
)

# config_request should be the first in the stream of requests.
config_request = videointelligence.StreamingAnnotateVideoRequest(
    video_config=video_config
)

# Set the chunk size to 5MB (recommended less than 10MB).
chunk_size = 5 * 1024 * 1024

def stream_generator():
    yield config_request
    # Load file content.
    # Note: Input videos must have supported video codecs. See
    # https://cloud.google.com/video-intelligence/docs/streaming/streaming#supported_video_codecs
    # for more details.
    with io.open(path, "rb") as video_file:
        while True:
            data = video_file.read(chunk_size)
            if not data:
                break
            yield videointelligence.StreamingAnnotateVideoRequest(
                input_content=data
            )

requests = stream_generator()

# streaming_annotate_video returns a generator.
# The default timeout is about 300 seconds.
# To process longer videos it should be set to
# larger than the length (in seconds) of the video.
responses = client.streaming_annotate_video(requests, timeout=900)

# Each response corresponds to about 1 second of video.
for response in responses:
    # Check for errors.
    if response.error.message:
        print(response.error.message)
        break

    for label in response.annotation_results.label_annotations:
        for frame in label.frames:
            print(
                "At {:3d}s segment, {:5.1%} {}".format(
                    frame.time_offset.seconds,
                    frame.confidence,
                    label.entity.entity_id,
                )
            )