Detect Frames (Activity)

This article is about the current version of Grooper. Note that some content may still need to be updated.
2025	2.80

view_module Detect Frames is a specialized Activity for processing microfiche in Grooper. It locates and identifies frame lines on microfiche card images, enabling the isolation of areas within the frames for further data extraction or processing.

The Detect Frames activity is a core component of Grooper’s microfiche processing workflow. Its primary function is to automatically locate and define the boundaries of individual document frames within each strip of a microfiche card, enabling accurate extraction and downstream processing of each page.

• Grooper's microfiche processing components were developed for and tested using Mekel brand scanners. Please contact support (support@bisok.com) to verify compatibility with other microfiche scanners.

What is Detect Frames?

Detect Frames is a specialized activity that analyzes the organized image tiles of a microfiche strip (prepared by the Initialize Card activity) to identify the precise locations of document frames. Each frame typically represents a single page or document stored on the microfiche card.

What is it for?

The main purposes of the Detect Frames activity are:

• Automatic frame detection: Identifies the grid of frames (pages) present on each strip of the fiche card, even in the presence of missing or low-quality frames.
• Quality assurance: Flags strips or frames that may require human review due to detection issues, such as missing frames or low image intensity.
• Preparation for extraction: Produces frame location data that is used by the Clip Frames activity to crop out each page as a separate image.

How does it work?

The Detect Frames activity performs several key operations:

1. Preview image generation:
   • Assembles a low-resolution preview of the strip from the sorted tiles, providing a basis for fast and efficient frame detection.
2. Image preprocessing:
   • Applies binarization and filtering to enhance contrast and reduce noise, making frame boundaries more distinct.
3. Gutter and grid detection:
   • Uses configurable properties such as "Minimum Vertical Length," "Minimum Horizontal Length," "Maximum Gap," and "Minimum Thickness" to detect the gutters (spaces) between frames.
   • Identifies the grid structure of the frames based on these gutters.
4. Frame validation and adjustment:
   • Validates each detected frame using properties like "Page Size Range" and "Minimum Intensity."
   • Flags frames that are too small, too dark, or otherwise problematic.
   • Optionally infers the positions of missing frames if "Infer Missing Frames" is enabled.
5. Diagnostics and flagging:
   • Generates diagnostic images and logs for review and troubleshooting.
   • Flags the Batch Folder if the number of detected frames does not match expectations or if frames require review.
6. Output:
   • Saves the detected frame information to a file ("FicheStripData.json").
   • This is used by the "Fiche Strip Viewer" to review frame detection.
   • This is used by the Clip Frames activity to crop out each detected frame and save it as a separate page.

General configuration

The Detect Frames activity is typically used as the second step in a microfiche Batch Process, following Initialize Card. To use it:

1. Add the Detect Frames activity to your Batch Process after Initialize Card.
   • Initialize Card must run first before Detect Frames can operate correctly.
2. Configure the following key properties to match your fiche card layout and image quality:
   • "Card Layout": Defines the expected arrangement of frames.
   • "Processing Resolution": Sets the DPI for frame detection and preview generation.
   • "Binarization": Controls how the image is converted to black and white for analysis.
   • "Minimum Vertical Length", "Minimum Horizontal Length", "Maximum Gap", "Minimum Thickness": Fine-tune gutter detection.
   • "Page Size Range": Sets the expected size range for valid frames.
   • "Minimum Intensity": Flags frames that are too dark.
   • "Maximum Empty Cells": Allows for a certain number of empty cells at the end of a card.
   • "Infer Missing Frames": Enables automatic inference of missing frame positions.

Example workflow

After Detect Frames completes, the frame boundaries are stored to each tile. The next step is typically Clip Frames which crops out each detected frame as a separate page for further processing.

A typical microfiche processing workflow might look like this:

1. Fiche scan files are imported into Grooper.
2. Initialize Card – Organize tiles and create a preview image.
3. Detect Frames – Detect frame locations on each strip.
4. Clip Frames – Extract each frame as a separate page.
5. Additional steps such as image cleanup, OCR, classification, and data extraction.

Best practices

• Carefully configure the "Card Layout" and detection parameters to match your specific fiche card format.
• Use the diagnostic images and logs to review detection accuracy and adjust settings as needed.
• Enable "Infer Missing Frames" if your fiche cards may have missing or damaged frames.
• Review flagged strips and frames using the Fiche Strip Viewer or other review tools before proceeding to extraction.

Related concepts

• Microfiche Processing
• Initialize Card
• Clip Frames