Region of Interest (ROI)

Barcode Recognition in a Region of Interest (ROI)

BarCodeReader in Aspose.BarCode for Java allows you to restrict recognition to specific rectangular areas of an image,
known as regions of interest (ROIs). This can significantly reduce processing time and improve recognition accuracy
when barcodes are located in known zones on complex layouts.

Typical scenarios for using ROI:

  • A barcode occupies a fixed area of a label or form.
  • Several different barcodes are placed in distinct regions of a single image.
  • You want to focus recognition on 1D or 2D barcodes in a specific part of the image.
  • You want to compare full-image recognition with targeted recognition inside a sub-region.
  • A barcode is offset inside a larger canvas (for example, centered with margins).

All examples in this article are based on the sample class:

com.aspose.barcode.guide.recognition.roi.RegionOfInterestExamples

You can find the full source code on GitHub:

RegionOfInterestExamples.java

In the snippets below, getFullPath("...") represents a helper that returns the file system path to the sample image.
In your code you can pass a plain file path string instead.

1. Limiting Recognition to a Single Region

The simplest use case is to scan only one rectangular area of the image that contains a specific barcode.
This helps to ignore all other content (background graphics, text, additional symbols) outside the region.

Rectangle roi = new Rectangle(20, 20, 440, 140);
BarCodeReader reader = new BarCodeReader(getFullPath("roi_canvas.png"), roi, DecodeType.CODE_128);

In this example:

  • roi defines a rectangular region where a Code 128 barcode is located.
  • The BarCodeReader constructor receives:
    • the image path,
    • the ROI rectangle,
    • the expected barcode type (DecodeType.CODE_128).

Only the specified region is scanned instead of the full image.

2. Using ROI with 2D Barcode Type Groups

If you know that a given region contains a 2D barcode, but do not want to hard-code a specific symbology,
you can pass the TYPES_2D group as the decode type. The ROI will still limit recognition to the selected area.

Rectangle roi = new Rectangle(470, 10, 250, 250);
BarCodeReader reader = new BarCodeReader(getFullPath("roi_canvas.png"), roi, DecodeType.TYPES_2D);

Here:

  • The rectangle bounds the area of a QR code on the sample canvas.
  • DecodeType.TYPES_2D instructs BarCodeReader to search for any supported 2D barcode type inside that region.
  • This approach is useful when the image layout is fixed, but the exact 2D symbology may vary.

3. Multiple ROIs in a Single Pass

A common pattern is to read several barcodes from fixed positions on the same image (for example, top-left and bottom-right corners).
BarCodeReader can accept an array of rectangles to scan multiple ROIs in one pass.

Rectangle[] areas = new Rectangle[] {
    new Rectangle(20, 20, 440, 140),
    new Rectangle(470, 240, 450, 160)
};

BarCodeReader reader = new BarCodeReader(getFullPath("roi_canvas.png"), areas, DecodeType.TYPES_1D);
BarCodeResult[] results = reader.readBarCodes();

In this example:

  • The first rectangle covers a Code 128 area.
  • The second rectangle covers an EAN-13 area.
  • DecodeType.TYPES_1D limits recognition to all supported 1D barcode types.
  • readBarCodes() returns all barcodes found in the specified ROIs.

This pattern is well suited for structured documents and labels with multiple fixed barcode fields.

4. Full-Image Scan vs ROI-Based Scan

Sometimes it is useful to first scan the entire image to confirm that multiple barcodes are present,
and then perform ROI-based recognition to focus on a particular symbology in a known area.

BarCodeReader fullReader =
        new BarCodeReader(getFullPath("roi_canvas.png"), DecodeType.ALL_SUPPORTED_TYPES);
BarCodeResult[] fullResults = fullReader.readBarCodes();

Rectangle dataMatrixArea = new Rectangle(10, 190, 230, 230);
BarCodeReader roiReader =
        new BarCodeReader(getFullPath("roi_canvas.png"), dataMatrixArea, DecodeType.DATA_MATRIX);
BarCodeResult[] roiResults = roiReader.readBarCodes();

Here:

  • The full-image reader uses DecodeType.ALL_SUPPORTED_TYPES to detect all possible barcodes on the canvas.
  • The ROI reader is restricted to a rectangle that contains a Data Matrix symbol and uses DecodeType.DATA_MATRIX.
  • This combination lets you:
    • verify the overall content of the image,
    • and then focus on a specific region and symbology for more precise processing.

5. ROI on an Image with an Offset Barcode

Barcodes are not always aligned with the image origin. They may be placed with margins or centered on a larger canvas.
In such cases, you can select a sub-region that encloses the symbol and ignore the rest of the image.

Rectangle roi = new Rectangle(150, 80, 480, 160);
BarCodeReader reader = new BarCodeReader(getFullPath("code128_offset.png"), roi, DecodeType.CODE_128);
BarCodeResult[] results = reader.readBarCodes();

In this scenario:

  • The file code128_offset.png contains a single Code 128 symbol drawn away from the top-left corner.
  • The ROI rectangle covers only the barcode area.
  • BarCodeReader processes only this sub-region, which is helpful when the margins or background graphics are not relevant.

Summary

Regions of interest (ROIs) allow you to control exactly where BarCodeReader looks for barcodes inside an image.
By providing one or more Rectangle instances to the constructor, you can:

  • Speed up recognition on large or complex images.
  • Reduce the influence of background noise and unrelated content.
  • Combine ROIs with type groups like TYPES_1D and TYPES_2D.
  • Compare full-image scanning with focused recognition in specific zones.
  • Work effectively with layouts where barcodes are placed away from the image origin.

All ROI examples shown here are based on:

RegionOfInterestExamples.java

Use the demonstrated patterns as a reference when adding ROI-based recognition to your own Java applications.

Barcode Properties Performance