IntelligentOCR アクティビティ

With Intelligent OCR activities you can process documents in a comprehensive manner, allowing you to not only digitize, extract, classify, and validate documents, but also train your extractor and classifiers on your specific data, so they can be faster and more accurate. The steps involved in creating Document Understanding^TM processes using Intelligent OCR activities are:

1. Create the Taxonomy: Define document types and convert them into a Document Object Model variable using the Load Taxonomy activity.
2. Digitize documents: Prepare documents so robots can process them using an OCR engine, by storing their text inside a String variable, and basic information about them inside a Document Object Model file.
3. Classify documents: Prepare documents using certain classifiers, so robots can identify what types of files they're processing.
4. Validate the classification of documents: Verify and validate that the documents have been correctly classified.
5. Train your classifiers: Configure your classifiers based on input received while validating the classification
6. Extract data from documents: Identify and extract specific information from your documents using various extractors to send it for validation.
7. Validate the extractions documents: Verify and validate the documents you processed, classified, and extracted, using the input of your team members within Action Center.
8. Train your extractors: Configure your extractors based on input received while validating the extraction.
9. Consume exported data: Once you validate the extracted data, you can use it as it is or export it as a DataSet variable using the Export Extraction Results activity.

はじめる前に

IntelligentOCR.Activities の使用を開始する前に、以下の特性を確認してください。

• 構成可能性が高い。そのため習得にも時間がかかります。
• 複数のオブジェクトとアクティビティが存在し、柔軟性を提供するように設計されている。
• 以下のような複雑な要因のために再利用性が低くなる。
  • ワークフロー内でさまざまな設定が必要である。
  • 以下のように、明示的な引数をあるアクティビティから別のアクティビティに何度も渡す必要がある。
    • タクソノミー
    • Document Object Model
    • テキスト
    • 分類結果
    • 抽出結果