Add possibility to set image preprocessing properties

PDFOC-102

Author: alexander.kischuk

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/ImagePreprocessingOptions.java

@@ -0,0 +1,110 @@
+/*
+    This file is part of the iText (R) project.
+    Copyright (c) 1998-2020 iText Group NV
+    Authors: iText Software.
+
+    This program is offered under a commercial and under the AGPL license.
+    For commercial licensing, contact us at https://itextpdf.com/sales.  For AGPL licensing, see below.
+
+    AGPL licensing:
+    This program is free software: you can redistribute it and/or modify
+    it under the terms of the GNU Affero General Public License as published by
+    the Free Software Foundation, either version 3 of the License, or
+    (at your option) any later version.
+
+    This program is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU Affero General Public License for more details.
+
+    You should have received a copy of the GNU Affero General Public License
+    along with this program.  If not, see <https://www.gnu.org/licenses/>.
+ */
+package com.itextpdf.pdfocr.tesseract4;
+
+/**
+ * Additional options applied on image preprocessing step.
+ */
+public class ImagePreprocessingOptions {
+
+    /**
+     * Adaptive threshold tile width as described here: http://www.leptonica.org/binarization.html.
+     * Default value of 0 is considered as full image width which means no tiling.
+     */
+    private int tileWidth;
+
+
+    /**
+     * Adaptive threshold tile height as described here: http://www.leptonica.org/binarization.html.
+     * Default value of 0 is considered as full image height which means no tiling.
+     */
+    private int tileHeight;
+
+    /**
+     * Adaptive threshold smoothing as described here: http://www.leptonica.org/binarization.html.
+     */
+    private boolean smoothTiling = true;
+
+    public ImagePreprocessingOptions() {
+    }
+
+    public ImagePreprocessingOptions(ImagePreprocessingOptions imagePreprocessingOptions) {
+        this.tileWidth = imagePreprocessingOptions.tileWidth;
+        this.tileHeight = imagePreprocessingOptions.tileHeight;
+        this.smoothTiling = imagePreprocessingOptions.smoothTiling;
+    }
+
+    /**
+     * Gets {@link #tileWidth}.
+     * @return tile width
+     */
+    final public int getTileWidth() {
+        return tileWidth;
+    }
+
+    /**
+     * Sets {@link #tileWidth}.
+     * @param tileWidth tile width
+     * @return {@link ImagePreprocessingOptions}
+     */
+    final public ImagePreprocessingOptions setTileWidth(int tileWidth) {
+        this.tileWidth = tileWidth;
+        return this;
+    }
+
+    /**
+     * Gets {@link #tileHeight}.
+     * @return tile height
+     */
+    final public int getTileHeight() {
+        return tileHeight;
+    }
+
+    /**
+     * Sets {@link #tileHeight}.
+     * @param tileHeight tile height
+     * @return {@link ImagePreprocessingOptions}
+     */
+    final public ImagePreprocessingOptions setTileHeight(int tileHeight) {
+        this.tileHeight = tileHeight;
+        return this;
+    }
+
+    /**
+     * Gets {@link #smoothTiling}.
+     * @return smooth tiling flag
+     */
+    final public boolean isSmoothTiling() {
+        return smoothTiling;
+    }
+
+    /**
+     * Sets {@link #smoothTiling}.
+     * @param smoothTiling smooth tiling flag
+     * @return {@link ImagePreprocessingOptions}
+     */
+    final public ImagePreprocessingOptions setSmoothTiling(boolean smoothTiling) {
+        this.smoothTiling = smoothTiling;
+        return this;
+    }
+}

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/ImagePreprocessingUtil.java

@@ -142,14 +142,17 @@ static BufferedImage readAsPixAndConvertToBufferedImage(
      * Performs basic image preprocessing using buffered image (if provided).
      * Preprocessed image will be saved in temporary directory.
      *
-     * @param inputFile input image {@link java.io.File}
+     * @param inputFile input image {@link File}
      * @param pageNumber number of page to be preprocessed
+     * @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
      * @return created preprocessed image as {@link net.sourceforge.lept4j.Pix}
      * @throws Tesseract4OcrException if it was not possible to read or convert
      * input file
      */
     static Pix preprocessImage(final File inputFile,
-            final int pageNumber) throws Tesseract4OcrException {
+                               final int pageNumber,
+                               final ImagePreprocessingOptions imagePreprocessingOptions)
+            throws Tesseract4OcrException {
         Pix pix = null;
         // read image
         if (isTiffImage(inputFile)) {
@@ -163,7 +166,7 @@ static Pix preprocessImage(final File inputFile,
                     Tesseract4OcrException.CANNOT_READ_PROVIDED_IMAGE)
                     .setMessageParams(inputFile.getAbsolutePath());
         }
-        return TesseractOcrUtil.preprocessPix(pix);
+        return TesseractOcrUtil.preprocessPix(pix, imagePreprocessingOptions);
     }
 
     /**

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/Tesseract4ExecutableOcrEngine.java

@@ -366,8 +366,9 @@ private String preprocessImage(final File inputImage,
         try {
             if (getTesseract4OcrEngineProperties().isPreprocessingImages()) {
                 Pix pix = ImagePreprocessingUtil
-                        .preprocessImage(inputImage, pageNumber);
-                TesseractOcrUtil.savePixToTempPngFile(tmpFileName, pix);
+                        .preprocessImage(inputImage, pageNumber,
+                                getTesseract4OcrEngineProperties().getImagePreprocessingOptions());
+                TesseractOcrUtil.savePixToPngFile(tmpFileName, pix);
                 if (!Files.exists(Paths.get(tmpFileName))) {
                     BufferedImage img = TesseractOcrUtil.convertPixToImage(pix);
                     if (img != null) {

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/Tesseract4LibOcrEngine.java

@@ -283,7 +283,8 @@ private String getOcrResultForSinglePage(final File inputImage,
                 result = new TesseractOcrUtil().getOcrResultAsString(
                         getTesseractInstance(),
                         ImagePreprocessingUtil
-                                .preprocessImage(inputImage, pageNumber),
+                                .preprocessImage(inputImage, pageNumber,
+                                        getTesseract4OcrEngineProperties().getImagePreprocessingOptions()),
                         outputFormat);
             }
             if (result == null) {

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/Tesseract4OcrEngineProperties.java

@@ -66,12 +66,6 @@ public class Tesseract4OcrEngineProperties extends OcrEngineProperties {
      */
     private Integer pageSegMode = 3;
 
-    /**
-     * "True" - if images need to be preprocessed, otherwise - false.
-     * True by default.
-     */
-    private boolean preprocessingImages = true;
-
     /**
      * Defines the way text is retrieved from tesseract output.
      * Default text positioning is by lines.
@@ -90,6 +84,11 @@ public class Tesseract4OcrEngineProperties extends OcrEngineProperties {
      */
     private boolean isUserWordsFileTemporary = false;
 
+    /**
+     * Settings for image preprocessing.
+     */
+    private ImagePreprocessingOptions imagePreprocessingOptions = new ImagePreprocessingOptions();
+
     /**
      * Creates a new {@link Tesseract4OcrEngineProperties} instance.
      */
@@ -107,9 +106,9 @@ public Tesseract4OcrEngineProperties(Tesseract4OcrEngineProperties other) {
         super(other);
         this.tessDataDir = other.tessDataDir;
         this.pageSegMode = other.pageSegMode;
-        this.preprocessingImages = other.preprocessingImages;
         this.textPositioning = other.textPositioning;
         this.pathToUserWordsFile = other.pathToUserWordsFile;
+        this.imagePreprocessingOptions = other.imagePreprocessingOptions;
     }
 
     /**
@@ -194,7 +193,7 @@ public final Tesseract4OcrEngineProperties setPageSegMode(
      * @return true if images need to be preprocessed, otherwise - false
      */
     public final boolean isPreprocessingImages() {
-        return preprocessingImages;
+        return imagePreprocessingOptions != null;
     }
 
     /**
@@ -206,7 +205,13 @@ public final boolean isPreprocessingImages() {
      */
     public final Tesseract4OcrEngineProperties setPreprocessingImages(
             final boolean preprocess) {
-        preprocessingImages = preprocess;
+        if (preprocess) {
+            if (imagePreprocessingOptions == null) {
+                imagePreprocessingOptions = new ImagePreprocessingOptions();
+            }
+        } else {
+            imagePreprocessingOptions = null;
+        }
         return this;
     }
 
@@ -387,4 +392,25 @@ final boolean isUserWordsFileTemporary() {
         return isUserWordsFileTemporary;
     }
 
+    /**
+     * Gets {@link #imagePreprocessingOptions}.
+     * @return {@link ImagePreprocessingOptions}
+     */
+    final public ImagePreprocessingOptions getImagePreprocessingOptions() {
+        return imagePreprocessingOptions;
+    }
+
+    /**
+     * Sets {@link #imagePreprocessingOptions}.
+     * @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
+     * @return the {@link Tesseract4OcrEngineProperties} instance
+     */
+    final public Tesseract4OcrEngineProperties setImagePreprocessingOptions(
+            ImagePreprocessingOptions imagePreprocessingOptions) {
+        this.imagePreprocessingOptions = imagePreprocessingOptions;
+        return this;
+    }
+
 }
+
+

pdfocr-tesseract4/src/main/java/com/itextpdf/pdfocr/tesseract4/TesseractOcrUtil.java

@@ -126,13 +126,15 @@ static Pix readPixPageFromTiff(final File inputFile,
      * converting to grayscale,
      * thresholding.
      *
-     * @param pix {@link net.sourceforge.lept4j.Pix} object to be processed
+     * @param pix {@link Pix} object to be processed
+     * @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
      * @return preprocessed {@link net.sourceforge.lept4j.Pix} object
      */
-    static Pix preprocessPix(Pix pix) {
-        pix = convertToGrayscale(pix);
-        pix = otsuImageThresholding(pix);
-        return pix;
+    static Pix preprocessPix(final Pix pix,
+                             final ImagePreprocessingOptions imagePreprocessingOptions) {
+        Pix pix1 = convertToGrayscale(pix);
+        pix1 = otsuImageThresholding(pix1, imagePreprocessingOptions);
+        return pix1;
     }
 
     /**
@@ -165,18 +167,27 @@ static Pix convertToGrayscale(final Pix pix) {
      * {@link net.sourceforge.lept4j.Leptonica#pixOtsuAdaptiveThreshold}
      * method.
      *
-     * @param pix {@link net.sourceforge.lept4j.Pix} object to be processed
+     * @param pix {@link Pix} object to be processed
+     * @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
      * @return {@link net.sourceforge.lept4j.Pix} object after thresholding
      */
-    static Pix otsuImageThresholding(final Pix pix) {
+    static Pix otsuImageThresholding(final Pix pix,
+                                     final ImagePreprocessingOptions imagePreprocessingOptions) {
         if (pix != null) {
             Pix thresholdPix = null;
             if (pix.d == 8) {
                 PointerByReference pointer = new PointerByReference();
                 Leptonica.INSTANCE
-                        .pixOtsuAdaptiveThreshold(pix, pix.w, pix.h,
-                                0, 0, 0,
-                                null, pointer);
+                        .pixOtsuAdaptiveThreshold(pix,
+                                getOtsuAdaptiveThresholdTileSize(pix.w,
+                                        imagePreprocessingOptions.getTileWidth()),
+                                getOtsuAdaptiveThresholdTileSize(pix.h,
+                                        imagePreprocessingOptions.getTileHeight()),
+                                getOtsuAdaptiveThresholdSmoothingTileSize(pix.w,
+                                        imagePreprocessingOptions.isSmoothTiling()),
+                                getOtsuAdaptiveThresholdSmoothingTileSize(pix.h,
+                                        imagePreprocessingOptions.isSmoothTiling()),
+                                0,null, pointer);
                 thresholdPix = new Pix(pointer.getValue());
                 if (thresholdPix.w > 0 && thresholdPix.h > 0) {
                     // destroying original pix
@@ -201,6 +212,36 @@ static Pix otsuImageThresholding(final Pix pix) {
         }
     }
 
+    /**
+     * Gets adaptive threshold tile size.
+     */
+    static int getOtsuAdaptiveThresholdTileSize(int imageSize, int tileSize) {
+        if (tileSize == 0) {
+            return imageSize;
+        } else {
+            return tileSize;
+        }
+    }
+
+    /**
+     * Gets adaptive threshold smoothing tile size.
+     * Can be either equal to page size or 0.
+     */
+    static int getOtsuAdaptiveThresholdSmoothingTileSize(int imageSize, boolean smoothTiling) {
+        if (smoothTiling) {
+            return imageSize;
+        } else {
+            return 0;
+        }
+    }
+
+    /**
+     * Gets an integer pixel in the default RGB color model.
+     */
+    static int getImagePixelColor(BufferedImage image, int x, int y) {
+        return image.getRGB(x, y);
+    }
+
     /**
      * Destroys {@link net.sourceforge.lept4j.Pix} object.
      *
@@ -403,15 +444,15 @@ static void saveImageToTempPngFile(final String tmpFileName,
     /**
      * Saves passed {@link net.sourceforge.lept4j.Pix} to given path
      *
-     * @param tmpFileName provided file path to save the
+     * @param filename provided file path to save the
      * {@link net.sourceforge.lept4j.Pix}
      * @param pix provided {@link net.sourceforge.lept4j.Pix} to be saved
      */
-    static void savePixToTempPngFile(final String tmpFileName,
-            final Pix pix) {
+    static void savePixToPngFile(final String filename,
+                                 final Pix pix) {
         if (pix != null) {
             try {
-                Leptonica.INSTANCE.pixWritePng(tmpFileName, pix,
+                Leptonica.INSTANCE.pixWritePng(filename, pix,
                         ILeptonica.IFF_PNG);
             } catch (Exception e) { // NOSONAR
                 LOGGER.info(MessageFormatUtil.format(