Machine Learning Platform for AI provides the video classification and object marking templates. When you create a labeling job, select a template that meets your requirements.

Video classification

Video classification is used to find one or more labels from a set of labels to match the content of an input video and attach the labels to the video. The template supports single-label and multi-label video classification.

  • Scenarios

    Video surveillance, live streaming recommendation, and short video recommendation.

  • Data structure
    • Input data
      Each row in the manifest file contains a topic. The topic must contain the videoUrl field.
      {"data":{"videoUrl":"oss://xxxxxxxxx.oss-cn-shanghai-internal.aliyuncs.com/video/English.mp4"}}
    • Output data
      Each row in the manifest file contains a topic and the labeling result. The following code provides an example of the JSON string in each row:
      {
        "data": {
          "videoUrl": "oss://xxxxxxxxxxx.oss-cn-shanghai-internal.aliyuncs.com/video/show.webm"
        },
        "label-bo0pc51ok7olwe****": {
          "results": [{
            "data": [{
              "data": ["Car promotion", "Promotion video"],
              "id": 165****,
              "type": "survey/multivalue"
            }],
            "id": "165****",
            "type": "survey"
          }]
        }
      }

Object marking

Object marking is used to locate and mark specific objects in a frame or specific frames of a video. The rectangle selection tool and polygon selection tool are commonly used.

  • Scenarios

    Autonomous driving, security surveillance, and video recommendation.

  • Data structure
    • Input data
      Each row in the manifest file contains a topic. The topic must contain the videoUrl field.
      {"data":{"videoUrl":"oss://xxxxxxxxx.oss-cn-shanghai-internal.aliyuncs.com/video/English.mp4"}}
    • Output data
      Each row in the manifest file contains a topic and the labeling results of one or more frames. The following code provides an example of the JSON string in each row:
      {
        "data": {
          "videoUrl": "oss://xxxxxxxxx.oss-cn-shanghai-internal.aliyuncs.com/video/English.mp4"
        },
        "label-rv0ih5l409u9x9****": {
          "results": [{
            "data": [{
              "speed_of_play": 1,
              "duration": 300.733375,
              "width": 960,
              "number_of_frames": 9022,
              "type": "video/meta",
              "frame_rate": 30,
              "height": 540
            }, {
              "frames": {
                "frame-443": {
                  "L6G-5Oc5STCSOuzKb****": {
                    "rotation": 0,
                    "x": 336.0483870967742,
                    "width": 488.4677419354839,
                    "y": 108.87096774193549,
                    "height": 349.83870967741933
                  }
                }
              },
              "type": "video/frames"
            }, {
              "frames": [443],
              "custom": {
                "Video type": "English learning",
                "Video name": "English learning",
                "Category": ["Education"]
              },
              "id": "L6G-5Oc5STCSOuzKb****",
              "type": "video/rectangleLabel",
              "value": "",
              "labelColor": "#ff7700"
            }],
            "id": "129****",
            "type": "video"
          }]
        }
      }