云愿景网络实体api与其演示之间的结果差异

时间:2018-03-30 07:34:18

标签: google-cloud-vision

当我调用云视觉网络实体检测时,当我在演示端(https://cloud.google.com/vision/)使用完全相同的图像时,我会得到不同的结果。

Api结果:

"webEntities": [
  {
    "entityId": "/m/069b9z",
    "score": 28.432,
    "description": "Muiden Castle"
  },
  {
    "entityId": "/m/0k3p",
    "score": 9.7216,
    "description": "Amsterdam"
  },
  {
    "entityId": "/m/0gvtk97",
    "score": 4.6464,
    "description": "Pampus"
  },
  {
    "entityId": "/m/03w10h",
    "score": 4.5328,
    "description": "Stelling van Amsterdam"
  },
  {
    "entityId": "/m/04rjz",
    "score": 0.7232,
    "description": "Middle Ages"
  }
],

演示结果:

 "webDetection": {
    "webEntities": [
      {
        "entityId": "/m/069b9z",
        "score": 28.2912,
        "description": "Muiden Castle"
      },
      {
        "entityId": "/m/06jsf",
        "score": 23.7056,
        "description": "Rijksmuseum"
      },
      {
        "entityId": "/m/03w10h",
        "score": 4.2744,
        "description": "Stelling van Amsterdam"
      },
      {
        "entityId": "/m/0gvtk97",
        "score": 4.2264,
        "description": "Pampus"
      },
      {
        "entityId": "/g/11clg9rz1_",
        "score": 1.3836,
        "description": "Marina Muiderzand"
      },
      {
        "entityId": "/m/0w0r9",
        "score": 1.1220801,
        "description": "Muiden"
      },
      {
        "entityId": "/m/0grl_",
        "score": 0.7032,
        "description": "Château"
      },
      {
        "entityId": "/m/0d5gx",
        "score": 0.5525,
        "description": "Castle"
      },
      {
        "entityId": "/m/01k12m",
        "score": 0.5256,
        "description": "Dutch Golden Age"
      },
      {
        "entityId": "/t/23p50h_kgbnhm",
        "score": 0.5227
      },
      {
        "entityId": "/m/026gg32",
        "score": 0.3787,
        "description": "Water castle"
      },
      {
        "entityId": "/g/122lvp4h",
        "score": 0.3709,
        "description": "Dag van het Kasteel"
      },
      {
        "entityId": "/m/0pgl9",
        "score": 0.3558,
        "description": "Tourist attraction"
      },
      {
        "entityId": "/m/0k3p",
        "score": 0.103136,
        "description": "Amsterdam"
      },
      {
        "entityId": "/m/059j2",
        "score": 2.0908006e-11,
        "description": "Netherlands"
      }
    ],

这就是我添加图片的方式

    using (Image image = Image.FromFile(filePath))
    {
        using (MemoryStream m = new MemoryStream())
        {
            image.Save(m, image.RawFormat);
            byte[] imageBytes = m.ToArray();

            request.Requests[0].Image.Content = Convert.ToBase64String(imageBytes);
        }
    }

几乎所有图像都会发生这种情况。

据我所知,您唯一的选择是,在调用Web实体时,检测api是您想要的结果量。但数量对结果的类型没有影响。

api的标签检测和地标检测给出了与演示相同的答案。

有人也见过这种行为吗? 我错过了其他选择吗?

修改 请求正文:

{
    "requests":[
    {
        "features":[
        {
            "type": "LABEL_DETECTION",
            "maxResults":5
        },
        {
            "type":"LANDMARK_DETECTION",
            "maxResults":1
        },
        {
            "type":"WEB_DETECTION",
            "maxResults":5
        }
        ],
        "image":
        {
            "content":"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"
        }
    }
    ] }

编辑2:
终点:https://vision.googleapis.com/v1/images:annotate
我不使用SDK。

编辑3:
我使用的图像:https://www.muiderslot.nl/wp-content/uploads/2015/05/DSC5093-Muiderslot-poster.jpg
奇怪的是,我刚刚再次测试了图像,现在演示还给出了第二个描述“阿姆斯特丹”,但第五个描述仍然不同。

但是这张图片仍然会产生更大的差异:https://imgur.com/a/r76Ah

0 个答案:

没有答案