Deconvolution Mode

Barcode Quality and Deconvolution Modes

Aspose.BarCode for Java provides flexible controls to balance recognition speed and robustness.
These controls are exposed through:

• QualitySettings presets (for example, getHighPerformance(), getNormalQuality(), getHighQuality())
• BarcodeQualityMode – selects recognition methods depending on the expected barcode quality (heavier methods for low-quality codes)
• DeconvolutionMode – deconvolution (image restoration) mode that defines the assumed level of image degradation and the strength of blur-reduction algorithms
• Targeted overrides such as XDimensionMode and setMinimalXDimension(...) (used together with XDimensionMode.USE_MINIMAL_X_DIMENSION when you need an explicit minimal module size)

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

com.aspose.barcode.guide.recognition.performance.BarcodeQualityDeconvolutionModeExample

You can find the full source code on GitHub:

DeconvolutionModeExample.java

In the snippets below, variables like imagePath represent paths to barcode images in your application.

1. DeconvolutionMode – blur and degradation handling

DeconvolutionMode is an image restoration mode that defines the assumed level of image degradation.
It controls how strong deblurring and related preprocessing algorithms are, and therefore how much extra time the engine spends on image restoration.

Supported values:

Mode	Description	Typical scenario
DeconvolutionMode.FAST	Lightweight restoration; minimal extra cost.	Slight blur or almost clean images.
DeconvolutionMode.NORMAL	Balanced restoration level.	Typical handheld photos with some blur/noise.
DeconvolutionMode.SLOW	Strongest restoration pipeline; highest CPU cost, maximum robustness.	Strong blur, motion blur, or heavy degradation.

Again, the pattern is:

1. Pick a QualitySettings preset.
2. Set the desired DeconvolutionMode.
3. Apply QualitySettings to BarCodeReader.

Example: configuring FAST, NORMAL, and SLOW for QR recognition.

String imagePath = ExampleAssist.pathCombine(FOLDER, "qset_qr.png");

// FAST: for clean or slightly blurred QR codes
BarCodeReader fastDeconvolutionReader = new BarCodeReader(imagePath, DecodeType.QR);
QualitySettings fastDeconvolutionSettings = QualitySettings.getHighQuality();
fastDeconvolutionSettings.setDeconvolution(DeconvolutionMode.FAST);
fastDeconvolutionReader.setQualitySettings(fastDeconvolutionSettings);

// NORMAL: default level for everyday mobile photos
BarCodeReader normalDeconvolutionReader = new BarCodeReader(imagePath, DecodeType.QR);
QualitySettings normalDeconvolutionSettings = QualitySettings.getNormalQuality();
normalDeconvolutionSettings.setDeconvolution(DeconvolutionMode.NORMAL);
normalDeconvolutionReader.setQualitySettings(normalDeconvolutionSettings);

// SLOW: strongest restoration for heavily blurred images
BarCodeReader slowDeconvolutionReader = new BarCodeReader(imagePath, DecodeType.QR);
QualitySettings slowDeconvolutionSettings = QualitySettings.getHighQuality();
slowDeconvolutionSettings.setDeconvolution(DeconvolutionMode.SLOW);
slowDeconvolutionReader.setQualitySettings(slowDeconvolutionSettings);

Use a higher deconvolution level only when necessary, because it increases processing time.

---

2. Combining presets with targeted overrides

A common pattern is to:

1. Start from a fast preset, such as getHighPerformance().
2. Apply targeted overrides for your specific scenario.

For example, if you expect small and low-quality 1D barcodes:

• enable detection of small modules via XDimensionMode.SMALL (heuristic mode, without MinimalXDimension),
• use BarcodeQualityMode.LOW and DeconvolutionMode.SLOW to improve robustness

Example: tuning recognition for small, degraded Code 128 symbols.

String imagePath = ExampleAssist.pathCombine(FOLDER, "qset_code128.png");

BarCodeReader barCodeReader = new BarCodeReader(imagePath, DecodeType.CODE_128);

QualitySettings qualitySettings = QualitySettings.getHighPerformance();
qualitySettings.setXDimension(XDimensionMode.SMALL);
qualitySettings.setBarcodeQuality(BarcodeQualityMode.LOW);
qualitySettings.setDeconvolution(DeconvolutionMode.SLOW);

barCodeReader.setQualitySettings(qualitySettings);

This configuration:

• keeps the overall preset tuned for performance,
• but adds extra robustness for tiny and degraded bars through explicit overrides.

---

3. QualitySettings presets overview

QualitySettings provides ready-made presets that enable appropriate internal algorithms for typical scenarios:

Preset method	Typical use case	Trade-off
QualitySettings.getHighPerformance()	Fast processing of clean, high-contrast images	Maximum speed, reduced robustness
QualitySettings.getNormalQuality()	General-purpose recognition for normal image quality	Balanced speed and robustness
QualitySettings.getHighQuality()	Difficult images: noise, blur, low contrast, tiny codes	Maximum robustness, lower performance

You can use these presets as a starting point and then adjust details with BarcodeQualityMode, DeconvolutionMode, and other properties.

---

4. BarcodeQualityMode – quality analysis profile

BarcodeQualityMode selects and enables recognition methods depending on the expected barcode quality.
For lower-quality barcodes it turns on heavier, more tolerant algorithms, which improves robustness but slows recognition.

Supported values:

Mode	Description	Typical scenario
BarcodeQualityMode.HIGH	Lightweight checks assuming good print quality and clear edges.	High-quality printed labels, synthetic test data.
BarcodeQualityMode.NORMAL	Standard analysis for regular-quality images.	Most business documents and mobile photos.
BarcodeQualityMode.LOW	Enables heavy and tolerant algorithms for damaged, low-contrast, or distorted barcodes.	Poor printing, fax/scans, degraded images.

The general pattern:

1. Choose a QualitySettings preset.
2. Set the desired BarcodeQualityMode.
3. Apply the configured QualitySettings to BarCodeReader.

Example: switching between HIGH, NORMAL, and LOW for Code 128.

String imagePath = ExampleAssist.pathCombine(FOLDER, "qset_code128.png");

// HIGH: fastest on clean images
BarCodeReader highQualityReader = new BarCodeReader(imagePath, DecodeType.CODE_128);
QualitySettings highPerformanceSettings = QualitySettings.getHighPerformance();
highPerformanceSettings.setBarcodeQuality(BarcodeQualityMode.HIGH);
highQualityReader.setQualitySettings(highPerformanceSettings);

// NORMAL: balanced default
BarCodeReader normalQualityReader = new BarCodeReader(imagePath, DecodeType.CODE_128);
QualitySettings normalQualitySettings = QualitySettings.getNormalQuality();
normalQualitySettings.setBarcodeQuality(BarcodeQualityMode.NORMAL);
normalQualityReader.setQualitySettings(normalQualitySettings);

// LOW: most tolerant, slowest
BarCodeReader lowQualityReader = new BarCodeReader(imagePath, DecodeType.CODE_128);
QualitySettings highQualitySettings = QualitySettings.getHighQuality();
highQualitySettings.setBarcodeQuality(BarcodeQualityMode.LOW);
lowQualityReader.setQualitySettings(highQualitySettings);

Use this pattern when you need to tune performance:

• Start with BarcodeQualityMode.HIGH if you expect consistently clean barcodes.
• Use BarcodeQualityMode.NORMAL as a general default.
• Switch to BarcodeQualityMode.LOW only for problematic inputs where robustness is more important than speed.

---

5. Practical guidelines

When configuring quality and deconvolution parameters:

• Start with presets:
  • getHighPerformance() for synthetic data or controlled environments.
  • getNormalQuality() for most real-world applications.
  • getHighQuality() for known difficult inputs.
• Use BarcodeQualityMode to match expected print quality:
  • HIGH – clean barcodes, best speed.
  • NORMAL – default choice if you are not sure.
  • LOW – damaged or low-contrast barcodes.
• Use DeconvolutionMode for blur and focus issues:
  • FAST – light restoration.
  • NORMAL – typical mobile snapshots.
  • SLOW – strong blur or motion artifacts.
• Combine presets with overrides (XDimensionMode, MinimalXDimension, etc.) when you know specific constraints of your data (for example, very small modules).

All patterns in this article are demonstrated in:

DeconvolutionModeExample.java

Use it as a reference when tuning recognition performance in your Java applications.