com.itextpdf.pdfocr.tesseract4.TesseractOcrUtil Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of pdfocr-tesseract4 Show documentation
Show all versions of pdfocr-tesseract4 Show documentation
pdfOCR-Tesseract4 is an iText add-on for Java to recognize and extract text in scanned documents and images. It can also convert them into fully ISO-compliant PDF or PDF/A-3u files that are accessible, searchable, and suitable for archiving
The newest version!
/*
This file is part of the iText (R) project.
Copyright (c) 1998-2024 Apryse Group NV
Authors: Apryse 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 imagePages =
Collections.emptyList();
/**
* Creates a new {@link TesseractOcrUtil} instance.
*/
TesseractOcrUtil() {
}
/**
* Reads required page from provided tiff image.
* Note that rotation is always applied when image read.
*
* @param inputFile input image as {@link java.io.File}
* @param pageNumber number of page
* @return result {@link net.sourceforge.lept4j.Pix} object created from
* given image
*/
static Pix readPixPageFromTiff(final File inputFile,
final int pageNumber) {
Pix pix = null;
BufferedImage img = TesseractOcrUtil
.getImagePage(inputFile, pageNumber);
if (img != null) {
pix = readPix(img);
}
// return required page to be preprocessed
return pix;
}
/**
* Performs default image preprocessing.
* It includes the following actions:
* converting to grayscale,
* thresholding.
*
* @param pix {@link Pix} object to be processed
* @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
* @return preprocessed {@link net.sourceforge.lept4j.Pix} object
*/
static Pix preprocessPix(final Pix pix,
final ImagePreprocessingOptions imagePreprocessingOptions) {
Pix pix1 = convertToGrayscale(pix);
pix1 = otsuImageThresholding(pix1, imagePreprocessingOptions);
return pix1;
}
/**
* Converts Leptonica {@link net.sourceforge.lept4j.Pix} to grayscale.
* In .Net image is converted only if this is 32bpp image. In java image is
* converted anyway using different Leptonica methods depending on
* image depth.
*
* @param pix {@link net.sourceforge.lept4j.Pix} object to be processed
* @return preprocessed {@link net.sourceforge.lept4j.Pix} object
*/
static Pix convertToGrayscale(final Pix pix) {
if (pix != null) {
int depth = LeptonicaWrapper.pixGetDepth(pix);
if (depth == 32) {
return LeptonicaWrapper.pixConvertRGBToLuminance(pix);
} else {
return LeptonicaWrapper.pixRemoveColormap(pix,
ILeptonica.REMOVE_CMAP_TO_GRAYSCALE);
}
} else {
return pix;
}
}
/**
* Performs Leptonica Otsu adaptive image thresholding using
* {@link net.sourceforge.lept4j.Leptonica#pixOtsuAdaptiveThreshold}
* method.
*
* @param inputPix {@link Pix} object to be processed
* @param imagePreprocessingOptions {@link ImagePreprocessingOptions}
* @return {@link net.sourceforge.lept4j.Pix} object after thresholding
*/
static Pix otsuImageThresholding(final Pix inputPix,
final ImagePreprocessingOptions imagePreprocessingOptions) {
if (inputPix != null) {
Pix binarizedPix = null;
if (inputPix.d == 8) {
PointerByReference pointer = new PointerByReference();
LeptonicaWrapper
.pixOtsuAdaptiveThreshold(inputPix,
getOtsuAdaptiveThresholdTileSize(inputPix.w,
imagePreprocessingOptions.getTileWidth()),
getOtsuAdaptiveThresholdTileSize(inputPix.h,
imagePreprocessingOptions.getTileHeight()),
getOtsuAdaptiveThresholdSmoothingTileSize(inputPix.w,
imagePreprocessingOptions.isSmoothTiling()),
getOtsuAdaptiveThresholdSmoothingTileSize(inputPix.h,
imagePreprocessingOptions.isSmoothTiling()),
0,null, pointer);
binarizedPix = new Pix(pointer.getValue());
if (binarizedPix.w > 0 && binarizedPix.h > 0) {
// destroying original pix
destroyPix(inputPix);
return binarizedPix;
} else {
final String logMessage = MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_BINARIZE_IMAGE,
inputPix.d);
LOGGER.info(logMessage);
// destroying created PointerByReference object
destroyPix(binarizedPix);
return inputPix;
}
} else {
LOGGER.info(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_BINARIZE_IMAGE,
inputPix.d));
return inputPix;
}
} else {
return inputPix;
}
}
/**
* Gets adaptive threshold tile size.
*/
static int getOtsuAdaptiveThresholdTileSize(int sizeOfImage, int sizeOfTile) {
if (sizeOfTile == 0) {
return sizeOfImage;
} else {
return sizeOfTile;
}
}
/**
* Gets adaptive threshold smoothing tile size.
* Can be either equal to page size or 0.
*/
static int getOtsuAdaptiveThresholdSmoothingTileSize(int sizeOfImage, boolean isSmoothTiling) {
if (isSmoothTiling) {
return sizeOfImage;
} 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.
*
* @param inputPix {@link net.sourceforge.lept4j.Pix} object to be destroyed
*/
static void destroyPix(Pix inputPix) {
if (inputPix != null) {
LeptonicaWrapper.lept_free(inputPix.getPointer());
}
}
/**
* Sets tesseract properties.
* The following properties are set in this method:
* In java: path to tess data, languages, psm
* In .Net: psm
* This means that other properties have been set during the
* initialization of tesseract instance previously or tesseract library
* doesn't provide such possibilities in api for .Net or java.
*
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract} object
* @param tessData path to tess data directory
* @param languages list of languages in required format
* as {@link java.lang.String}
* @param pageSegmentationMode page segmentation mode {@link java.lang.Integer}
* @param userWordsFilePath path to a temporary file with user words
*/
static void setTesseractProperties(
final ITesseract tesseractInstance,
final String tessData, final String languages,
final Integer pageSegmentationMode, final String userWordsFilePath) {
tesseractInstance.setDatapath(tessData);
tesseractInstance.setLanguage(languages);
if (pageSegmentationMode != null) {
tesseractInstance.setPageSegMode(pageSegmentationMode);
}
tesseractInstance.setOcrEngineMode(userWordsFilePath != null ? 0 : 3);
}
/**
* Creates tesseract instance with parameters.
* Method is used to initialize tesseract instance with parameters if it
* haven't been initialized yet.
* In this method in java 'tessData', 'languages' and 'userWordsFilePath'
* properties are unused as they will be set using setters in
* {@link #setTesseractProperties} method. In .Net all these properties
* are needed to be provided in tesseract constructor in order to
* initialize tesseract instance. Thus, tesseract initialization takes
* place in {@link Tesseract4LibOcrEngine#Tesseract4LibOcrEngine} constructor in
* java, but in .Net it happens only after all properties are validated,
* i.e. just before OCR process.
*
* @param isWindowsPlatform true is current os is windows
* @param tessData path to tess data directory
* @param languages list of languages in required format as
* {@link java.lang.String}
* @param userWordsFilePath path to a temporary file with user words
* @return initialized {@link net.sourceforge.tess4j.ITesseract} object
*/
static ITesseract initializeTesseractInstance(final boolean isWindowsPlatform,
final String tessData, final String languages,
final String userWordsFilePath) {
try {
if (isWindowsPlatform) {
return new Tesseract1();
} else {
return new Tesseract();
}
} catch (LinkageError e) {
throw new PdfOcrTesseract4Exception(isWindowsPlatform ?
PdfOcrTesseract4ExceptionMessageConstant.TESSERACT_LIB_NOT_INSTALLED_WIN :
PdfOcrTesseract4ExceptionMessageConstant.TESSERACT_LIB_NOT_INSTALLED, e);
}
}
/**
* Returns true if tesseract instance has been already disposed.
* (used in .net version)
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract}
* object to check
* @return true if tesseract instance is disposed.
*/
static boolean isTesseractInstanceDisposed(
final ITesseract tesseractInstance) {
return false;
}
/**
* Disposes {@link net.sourceforge.tess4j.ITesseract} instance.
* (used in .net version)
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract}
* object to dispose
*/
static void disposeTesseractInstance(
final ITesseract tesseractInstance) {
}
static Pix readPixFromFile(File inputImage) {
return LeptonicaWrapper.pixRead(inputImage.getAbsolutePath());
}
/**
* Converts Leptonica {@link net.sourceforge.lept4j.Pix}
* to {@link java.awt.image.BufferedImage} with
* {@link net.sourceforge.lept4j.ILeptonica#IFF_PNG} image format.
*
* @param inputPix input {@link net.sourceforge.lept4j.Pix} object
* @return result {@link java.awt.image.BufferedImage} object
* @throws IOException if it is not possible to convert
*/
static BufferedImage convertPixToImage(final Pix inputPix)
throws IOException {
if (inputPix != null) {
BufferedImage bi = null;
PointerByReference pdata = new PointerByReference();
try {
NativeSizeByReference psize = new NativeSizeByReference();
LeptonicaWrapper.pixWriteMem(pdata, psize, inputPix, ILeptonica.IFF_PNG);
byte[] b = pdata.getValue().getByteArray(0,
psize.getValue().intValue());
try (InputStream in = new ByteArrayInputStream(b)) {
bi = ImageIO.read(in);
}
} finally {
LeptonicaWrapper.lept_free(pdata.getValue());
}
return bi;
} else {
return null;
}
}
/**
* Gets path to temp file in current system temporary directory.
*
* @return path to temp file in the system temporary directory
*/
static String getTempFilePath(String name, String suffix) {
String tmpFileName = name + suffix;
try {
Path tempPath = Files.createTempFile(name, suffix);
tmpFileName = tempPath.toString();
} catch (IOException | IllegalArgumentException e) {
LOGGER.info(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_GET_TEMPORARY_DIRECTORY,
e.getMessage()));
}
return tmpFileName;
}
/**
* Gets requested image page from the provided image.
*
* @param inputFile input image
* @param page requested image page
* @return requested image page as a {@link java.awt.image.BufferedImage}
*/
static BufferedImage getImagePage(File inputFile, int page)
{
BufferedImage img = null;
try (InputStream is =
new FileInputStream(inputFile.getAbsolutePath())) {
List pages = Imaging.getAllBufferedImages(is,
inputFile.getAbsolutePath());
if (page >= pages.size()) {
LOGGER.warn(MessageFormatUtil.format(
Tesseract4LogMessageConstant.PAGE_NUMBER_IS_INCORRECT,
page,
inputFile.getAbsolutePath()));
return null;
}
img = pages.get(page);
} catch (ImageReadException | IOException e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant
.CANNOT_RETRIEVE_PAGES_FROM_IMAGE,
inputFile.getAbsolutePath(),
e.getMessage()));
}
return img;
}
/**
* Saves passed {@link java.awt.image.BufferedImage} to given path
*
* @param tmpFileName provided file path to save the
* {@link java.awt.image.BufferedImage}
* @param image provided {@link java.awt.image.BufferedImage} to be saved
*/
static void saveImageToTempPngFile(final String tmpFileName,
final BufferedImage image) {
if (image != null) {
try {
ImageIO.write(image, "png", new File(tmpFileName));
} catch (Exception e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_PROCESS_IMAGE,
e.getMessage()));
}
}
}
/**
* Saves passed {@link net.sourceforge.lept4j.Pix} to given path
*
* @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 savePixToPngFile(final String filename,
final Pix pix) {
if (pix != null) {
try {
LeptonicaWrapper.pixWritePng(filename, pix,
ILeptonica.IFF_PNG);
} catch (Exception e) {
LOGGER.info(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_PROCESS_IMAGE,
e.getMessage()));
}
}
}
/**
* Create temporary copy of input file to avoid issue with tesseract and
* different encodings in the path.
*
* @param src path to the source image
* @param dst destination path
*/
static void createTempFileCopy(final String src, final String dst)
throws IOException {
Files.copy(Paths.get(src), Paths.get(dst),
StandardCopyOption.REPLACE_EXISTING);
}
/**
* Returns parent directory for the passed path.
*
* @param path path to file
* @return parent directory where the file is located
*/
static String getParentDirectoryFile(final String path) {
return new File(path).getParent();
}
/**
* Retrieves list of pages from provided image as list of
* {@link java.awt.image.BufferedImage}, one per page and updates
* this list for the image using {@link #setListOfPages} method.
*
* @param inputFile input image {@link java.io.File}
*/
void initializeImagesListFromTiff(
final File inputFile) {
try (InputStream is =
new FileInputStream(inputFile.getAbsolutePath())) {
setListOfPages(Imaging
.getAllBufferedImages(is,
inputFile.getAbsolutePath()));
} catch (Exception e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant
.CANNOT_RETRIEVE_PAGES_FROM_IMAGE,
inputFile.getAbsolutePath(),
e.getMessage()));
}
}
/**
* Gets list of page of processing image as list of
* {@link java.awt.image.BufferedImage}, one per page.
*
* @return result {@link java.util.List} of pages
*/
List getListOfPages() {
return new ArrayList(imagePages);
}
/**
* Sets list of page of processing image as list of
* {@link java.awt.image.BufferedImage}, one per page.
*
* @param listOfPages list of {@link java.awt.image.BufferedImage} for
* each page.
*/
void setListOfPages(final List listOfPages) {
imagePages = Collections.unmodifiableList(listOfPages);
}
/**
* Performs ocr for the provided image
* and returns result as string in required format.
* ({@link OutputFormat} is used in .Net version,
* in java output format should already be set)
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract}
* object to perform OCR
* @param image input {@link java.awt.image.BufferedImage} to be processed
* @param outputFormat selected {@link OutputFormat} for tesseract
* @return result as {@link java.lang.String} in required format
* @throws TesseractException if tesseract recognition failed
*/
String getOcrResultAsString(
final ITesseract tesseractInstance,
final BufferedImage image, final OutputFormat outputFormat)
throws TesseractException {
if (image != null) {
return tesseractInstance.doOCR(image);
} else {
return null;
}
}
/**
* Performs ocr for the provided image
* and returns result as string in required format.
* ({@link OutputFormat} is used in .Net version, in java output format
* should already be set)
*
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract}
* object to perform OCR
* @param image input image as {@link java.io.File} to be
* processed
* @param outputFormat selected {@link OutputFormat} for tesseract
* @return result as {@link java.lang.String} in required format
* @throws TesseractException if tesseract recognition failed
*/
String getOcrResultAsString(
final ITesseract tesseractInstance,
final File image, final OutputFormat outputFormat)
throws TesseractException {
if (image != null) {
return tesseractInstance.doOCR(image);
} else {
return null;
}
}
/**
* Performs ocr for the provided image
* and returns result as string in required format.
* ({@link OutputFormat} is used in .Net version, in java output format
* should already be set)
*
* @param tesseractInstance {@link net.sourceforge.tess4j.ITesseract}
* object to perform OCR
* @param pix input image as {@link net.sourceforge.lept4j.Pix} to be
* processed
* @param outputFormat selected {@link OutputFormat} for tesseract
* @return result as {@link java.lang.String} in required format
* @throws TesseractException if tesseract recognition failed
* @throws IOException if it is not possible to convert input image
*/
String getOcrResultAsString(
final ITesseract tesseractInstance,
final Pix pix, final OutputFormat outputFormat)
throws TesseractException, IOException {
if (pix != null) {
BufferedImage bufferedImage = convertPixToImage(pix);
return getOcrResultAsString(tesseractInstance,
bufferedImage, outputFormat);
} else {
return null;
}
}
/**
* Read {@link net.sourceforge.lept4j.Pix} from {@link java.awt.image.BufferedImage}.
* Note that rotation is always applied when image read.
*
* @param image {@link java.awt.image.BufferedImage} to read from
* @return Pix result {@link net.sourceforge.lept4j.Pix}
*/
static Pix readPix(final BufferedImage image) {
try {
ByteArrayOutputStream baos = new ByteArrayOutputStream();
ImageIO.write(image, "png", baos);
return readPix(baos.toByteArray());
} catch (IOException e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_READ_INPUT_IMAGE,
e.getMessage()));
return null;
}
}
/**
* Read {@link net.sourceforge.lept4j.Pix} from {@link java.io.File}.
* Note that rotation is always applied when image read.
*
* @param inputFile {@link java.io.File} to read from
* @return Pix result {@link net.sourceforge.lept4j.Pix}
*/
static Pix readPix(final File inputFile) {
try {
return readPix(Files.readAllBytes(inputFile.toPath()));
} catch (IOException e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_READ_INPUT_IMAGE,
e.getMessage()));
return null;
}
}
/**
* Read {@link net.sourceforge.lept4j.Pix} from byte array.
* Note that rotation is always applied when image read.
*
* @param imageBytes to read from
* @return Pix result {@link net.sourceforge.lept4j.Pix}
*/
static Pix readPix(final byte[] imageBytes) {
Pix pix = null;
try {
ByteBuffer bb = ByteBuffer.wrap(imageBytes);
NativeSize size = new NativeSize(imageBytes.length);
pix = LeptonicaWrapper.pixReadMem(bb, size);
} catch (Exception e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_READ_INPUT_IMAGE,
e.getMessage()));
}
if (pix != null) {
int rotation = detectRotation(imageBytes);
pix = rotate(pix, rotation);
}
return pix;
}
/**
* Detect rotation specified by image metadata.
*
* @param file file to detect rotation
* @return image rotation as specified in metadata
*/
static int detectRotation(final File file) {
try {
return detectRotation(Files.readAllBytes(file.toPath()));
} catch (Exception e) {
LOGGER.error(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_READ_INPUT_IMAGE,
e.getMessage()));
return ROTATION_0;
}
}
/**
* Detect rotation specified by image metadata.
*
* @param imageData to detect rotation
* @return image rotation as specified in metadata
*/
static int detectRotation(final ImageData imageData) {
return detectRotation(imageData.getData());
}
/**
* Detect rotation specified by image metadata.
*
* @param data image data to detect rotation
* @return image rotation as specified in metadata
*/
static int detectRotation(final byte[] data) {
int rotation = ROTATION_0;
try {
final ImageMetadata metadata = Imaging.getMetadata(data);
TiffImageMetadata tiffImageMetadata;
if (metadata instanceof JpegImageMetadata) {
tiffImageMetadata = ((JpegImageMetadata) metadata).getExif();
} else if (metadata instanceof TiffImageMetadata) {
tiffImageMetadata = (TiffImageMetadata) metadata;
} else {
tiffImageMetadata = null;
}
if (tiffImageMetadata != null) {
rotation = readRotationFromMetadata(tiffImageMetadata);
}
} catch (Exception e) {
LOGGER.info(MessageFormatUtil.format(
Tesseract4LogMessageConstant.CANNOT_READ_IMAGE_METADATA,
e.getMessage()));
}
return rotation;
}
/**
* Reads image orientation from metadata and converts to rotation.
* @param tiffImageMetadata image metadata
* @return rotation
*/
static int readRotationFromMetadata(TiffImageMetadata tiffImageMetadata) {
final List items = tiffImageMetadata.getItems();
for (final Object item : items) {
if (item instanceof TiffImageMetadata.TiffMetadataItem &&
"Orientation".equals(((TiffImageMetadata.TiffMetadataItem) item).getKeyword())) {
int orientation = Integer.parseInt(((TiffImageMetadata.TiffMetadataItem) item).getText());
switch (orientation) {
case EXIF_ROTATION_0:
return ROTATION_0;
case EXIF_ROTATION_90:
return ROTATION_90;
case EXIF_ROTATION_180:
return ROTATION_180;
case EXIF_ROTATION_270:
return ROTATION_270;
default:
LOGGER.warn(MessageFormatUtil.format(
Tesseract4LogMessageConstant.UNSUPPORTED_EXIF_ORIENTATION_VALUE,
orientation));
return ROTATION_0;
}
}
}
return ROTATION_0;
}
/**
* Rotates image by specified angle.
*
* @param pix image source represented by {@link net.sourceforge.lept4j.Pix}
* @param rotation to rotate image at
* @return rotated image, if rotation differs from 0
*/
static Pix rotate(final Pix pix, int rotation) {
switch (rotation) {
case ROTATION_90:
return LeptonicaWrapper.pixRotate90(pix, 1);
case ROTATION_180:
return LeptonicaWrapper.pixRotate180(pix, pix);
case ROTATION_270:
return LeptonicaWrapper.pixRotate90(pix, -1);
default:
return pix;
}
}
/**
* Detects and applies rotation to image.
*
* @param imageData source image to rotate if needed
* @return rotated image, if rotation differs from 0
*/
static ImageData applyRotation(final ImageData imageData) {
Pix pix = readPix(imageData.getData());
if (pix == null) {
return imageData;
} else {
ImageData newImageData = imageData;
try {
PointerByReference data = new PointerByReference();
NativeSizeByReference size = new NativeSizeByReference();
if (LeptonicaWrapper.pixWriteMemPng(data, size, pix, 0) == 0) {
newImageData = ImageDataFactory.create(
data.getValue().getByteArray(0, size.getValue().intValue())
);
}
} finally {
destroyPix(pix);
}
return newImageData;
}
}
}
© 2015 - 2024 Weber Informatics LLC | Privacy Policy