Azure认知服务OCR给出不同的结果-如何补救? [英] Azure Cognitive Services OCR giving differing results - how to remedy?

问题描述

Azure CS在以下位置具有OCR演示(西向中枢端点)

Azure CS has an OCR demo (westcentralus endpoint) at

https://azure.microsoft.com/en-us/services/cognitive-services/computer-vision/?v = 18.05

在不良的测试图像上(恐怕我不能发布，因为它是身份证明文件)，我得到的OCR结果实际上与三个测试用例的实际文本都100％匹配-了不起.

On a poor test image (which I'm afraid I can't post because it's an identity document), I get OCR results that 100% match the actual text for three test cases in fact - remarkable.

但是，当我在下面的URL上使用Westeurope端点跟踪示例时，我得到的OCR结果更差-缺少一些文本:

However, when I follow the sample at the URL below, with the westeurope endpoint, I get poorer OCR results - some text is missing:

https://docs.microsoft.com/zh-CN/azure/cognitive-services/Computer-vision/quickstarts/python-print-text

这是为什么?更重要的是-如何访问v = 18.05端点?

Why is this? More to the point - how do I access the v=18.05 endpoint?

感谢所有快速的帮助.

推荐答案

我想我明白了:您在提到的两个页面之间没有使用相同的操作.

I think I got your point: you are not using the same operation between the 2 pages you mention.

如果您阅读的是工作演示上方的段落，则表示

If you read the paragraph just above the working demo you are mentioning here it says:

开始全面使用OCR服务，并发现下面的新预览OCR引擎(通过识别"文字"API操作)，具有更好的文本识别结果，英语.

Get started with the OCR service in general availability, and discover below a sneak peek of the new preview OCR engine (through "Recognize Text" API operation) with even better text recognition results for English.

如果您查看的是其他文档(

And if you have a look to the other documentation you are pointing at (this one), they are using the OCR operation:

vision_base_url = "https://westcentralus.api.cognitive.microsoft.com/vision/v2.0/"

ocr_url = vision_base_url + "ocr"

因此，如果要使用此新的预览版本，请将操作更改为 recognizeText

So if you want to use this new preview version, change the operation to recognizeText

它在西欧地区可用(请参见)，然后我进行了一个快速测试:Azure演示页上提供的示例正在使用此操作，而不是另一个.

It is available in West Europe region (see here), and I made a quick test: the samples provided on Azure demo page are working with this operation, and not in the other one.

但这一次该操作需要2个调用:

But this time the operation needs 2 calls:

• 一个POST操作来提交您的请求( recognizeText 操作)，在那里您将得到一个 202 Accepted 答案以及一个 operationId
• 一个GET操作符与上一步中的OperationId一起获得结果( textOperations 操作).例如: https://westeurope.api.cognitive.microsoft.com/vision/v2.0/textOperations/yourOperationId

• One POST operation to submit your request (recognizeText operation), where you will have a 202 Accepted answer with an operationId
• One GET opertaion to get the results (textOperations operation), with your OperationId from the previous step. For example: https://westeurope.api.cognitive.microsoft.com/vision/v2.0/textOperations/yourOperationId

对于来自Microsoft演示的CLOSED标志:

For the CLOSED sign from Microsoft Demos:

OCR操作的结果:

{
  "language": "unk",
  "orientation": "NotDetected",
  "textAngle": 0.0,
  "regions": []
}

带有RecognizeText的结果:

{
  "status": "Succeeded",
  "recognitionResult": {
    "lines": [{
      "boundingBox": [174, 488, 668, 675, 617, 810, 123, 622],
      "text": "CLOSED",
      "words": [{
        "boundingBox": [164, 494, 659, 673, 621, 810, 129, 628],
        "text": "CLOSED"
      }]
    }, {
      "boundingBox": [143, 641, 601, 811, 589, 843, 132, 673],
      "text": "WHEN ONE DOOR CLOSES, ANOTHER",
      "words": [{
        "boundingBox": [147, 646, 217, 671, 205, 698, 134, 669],
        "text": "WHEN"
      }, {
        "boundingBox": [230, 675, 281, 694, 269, 724, 218, 703],
        "text": "ONE"
      }, {
        "boundingBox": [291, 697, 359, 722, 348, 754, 279, 727],
        "text": "DOOR"
      }, {
        "boundingBox": [370, 726, 479, 767, 469, 798, 359, 758],
        "text": "CLOSES,"
      }, {
        "boundingBox": [476, 766, 598, 812, 588, 839, 466, 797],
        "text": "ANOTHER"
      }]
    }, {
      "boundingBox": [56, 668, 645, 886, 633, 919, 44, 700],
      "text": "OPENS.ALL YOU HAVE TO DO IS WALK IN",
      "words": [{
        "boundingBox": [74, 677, 223, 731, 213, 764, 65, 707],
        "text": "OPENS.ALL"
      }, {
        "boundingBox": [233, 735, 291, 756, 280, 789, 223, 767],
        "text": "YOU"
      }, {
        "boundingBox": [298, 759, 377, 788, 367, 821, 288, 792],
        "text": "HAVE"
      }, {
        "boundingBox": [387, 792, 423, 805, 413, 838, 376, 824],
        "text": "TO"
      }, {
        "boundingBox": [431, 808, 472, 824, 461, 855, 420, 841],
        "text": "DO"
      }, {
        "boundingBox": [479, 826, 510, 838, 499, 869, 468, 858],
        "text": "IS"
      }, {
        "boundingBox": [518, 841, 598, 872, 587, 901, 506, 872],
        "text": "WALK"
      }, {
        "boundingBox": [606, 875, 639, 887, 627, 916, 594, 904],
        "text": "IN"
      }]
    }]
  }
}

这篇关于Azure认知服务OCR给出不同的结果-如何补救?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！