Azure API及iText7/IronOCR能否提取手写问卷复选框数据？

Hey there! Let's walk through your questions with practical insights from working with document processing tools:

Good news—Azure Document Intelligence (formerly Form Recognizer) is exactly what you need here, and it integrates seamlessly with your C# stack. The older Azure Read API is great for free-form text, but Document Intelligence is built specifically for structured documents like your questionnaires.

• Its prebuilt "Form" model can automatically detect checkboxes, radio buttons, and their selection status. When you call the API, you'll get a response that includes properties like IsSelected for each checkbox. You can easily map this boolean value to your desired string (e.g., return "disagree" if the checkbox for Question #1 is marked as selected).
• For more tailored needs (like matching checkboxes to specific questions consistently), you can train a custom model using your own questionnaire templates—this boosts accuracy for your specific document layout.

If you need to step outside Azure, these options work well for checkbox extraction:

• AWS Textract: A robust document analysis service with native support for detecting checkboxes and their status. It can also link checkboxes to adjacent question text, making it easy to map selections to your database fields.
• Google Cloud Vision API: Its form parsing feature identifies checkboxes and returns their selection state. It’s strong for mixed document types (handwritten + printed + checkboxes).
• Tesseract OCR (Open Source): You’ll need to pair it with image processing libraries like OpenCV to preprocess scans (e.g., thresholding to highlight checkbox fills) and write custom logic to detect if a checkbox is checked. It’s free but requires more development effort than managed services.

Let’s break down these two C# libraries:

• IronOCR: This library has out-of-the-box support for checkbox detection. When you run OCR on your scanned documents, the OcrResult object includes a CheckBoxes collection—each entry has an IsChecked property you can use to map to your desired strings (like "disagree"). It handles scan preprocessing automatically, so it’s a low-effort option for C# projects.
• iText7: iText7 is primarily a PDF manipulation library, not a dedicated OCR tool. While it can integrate with Tesseract via its OCR module to extract text, detecting checkbox status requires custom work: you’ll need to first locate checkbox coordinates, then analyze pixel data or shape patterns to determine if it’s checked. It’s flexible but demands more code compared to IronOCR.

内容的提问来源于stack exchange，提问作者darego101