Advanced OCR¶
Advanced OCR gives you full control for high-accuracy text extraction from complex documents, scanned PDFs, and low-quality images. You can choose between Paddle OCR for AI-powered accuracy or Tesseract OCR for multi-language support, then fine-tune settings like confidence level, PDF processing mode, and display options for better results.
It also includes powerful Image Preprocessing tools such as denoising, sharpening, grayscale conversion, thresholding, deskew, and layout preservation. By configuring preprocessing steps and execution order, you can significantly improve text clarity and extraction accuracy -- perfect for professional and advanced OCR workflows.
When to Use Advanced OCR
Use Advanced OCR when Easy OCR doesn't give satisfactory results, especially for:
- Scanned documents with noise or low resolution
- Complex layouts with tables, columns, or mixed content
- Images taken at angles or with poor lighting
- Documents requiring specific preprocessing steps
Video Tutorial¶
Step-by-Step Guide¶
Step 1: Open Advanced OCR¶
From the Kaizen OCR dashboard, click on the Advanced OCR tile to open the Advanced OCR module.
Step 2: Add Your File¶
Click the file browser button or drag & drop your image or PDF into the application.
Step 3: Select Your File¶
Choose the document or image you want to process from the file dialog.
Step 4: Configure Settings and Start OCR¶
Configure your OCR settings:
- OCR Engine: Choose between Paddle OCR or Tesseract OCR
- Confidence Level: Adjust the minimum confidence threshold
- Preprocessing: Enable tools like denoising, sharpening, grayscale, thresholding, and deskew
- Execution Order: Arrange preprocessing steps in your preferred order
Once configured, click Start OCR to begin the extraction.
Step 5: View and Export Results¶
Review the extracted text in the results panel. The output will reflect the improved accuracy from your preprocessing settings.
Preprocessing Tools¶
| Tool | Description |
|---|---|
| Denoising | Removes noise and artifacts from scanned images |
| Sharpening | Enhances text edges for better recognition |
| Grayscale | Converts color images to grayscale for consistent processing |
| Thresholding | Converts images to black and white for clearer text |
| Deskew | Straightens tilted or rotated documents |
| Layout Preservation | Maintains the original document structure in output |