Train the Document Quality Analyzer on Your Data

The Document Quality Analyzer decides whether a user-provided document image is of sufficient quality to proceed, or whether the user should be prompted to capture a better one.

Distinguishing "good" from "bad" quality images is inherently difficult, because the definition of quality depends heavily on your use-case:

• Document type: What kinds of documents will be scanned? E.g. receipts, invoices, contracts, etc.
• Region of interest: Which part of the document carries the important information? For example, does the fine print need to be legible, or is it enough for the larger text to be readable?
• Downstream processing: What are the capabilities of the next processing passes? For instance, text with poor contrast may be fine for OCR but hard for humans to read.

By default, the Document Quality Analyzer is configured to work well across a wide range of use-cases. If your workflow has more specific requirements, you can use the Document Quality Analyzer Configurator to train the analyzer on your own dataset of document images.

The configurator learns from examples of accepted and rejected images that are representative of your use case, producing a custom configuration optimized for your workflow.

Using the Document Quality Analyzer Configurator

To create a custom configuration, provide a set of images representative of your use case (the "training data"), organized into two folders:

• Good quality images: Samples that should be accepted by the analyzer.
• Bad quality images: Samples that should be rejected by the analyzer, such as images that caused downstream processing failures or required users to retake the document.

The tool produces two outputs:

• A configuration file that you pass to the Document Quality Analyzer when initializing it.
• A training report that estimates expected performance and helps determine whether additional training samples may improve results.

In a typical scenario, around 100 images per class (good and bad) are sufficient to generate a reliable configuration.

Privacy and Deployment

The Document Quality Analyzer Configurator runs entirely within your environment. Training images remain under your control and do not need to be uploaded to Scanbot, Apryse, or any third-party service.

Availability

This capability is currently in closed beta.

Beta participants receive a free license key for the Document Quality Analyzer Configurator, along with onboarding guidance and direct support to help train and evaluate configurations for their specific use cases.

To request access, sign up here.

Want to scan longer than one minute?

Generate a free trial license to test the Scanbot SDK thoroughly.

Get free trial license

Scanbot SDK is part of the Apryse SDK product family

A mobile scan is just the start. With Apryse SDKs, you can expand mobile workflows into full cross‑platform document processing. Whether you need to edit PDFs, add secure digital signatures, or use a fast, customizable document viewer and editor, Apryse gives you the tools to build powerful features quickly.

Learn more

Was this helpful?