Package com.itextpdf.pdfocr.onnx.recognition

Class OnnxRecognitionPredictor

java.lang.Object

com.itextpdf.pdfocr.onnx.AbstractOnnxPredictor<BufferedImage, String>

com.itextpdf.pdfocr.onnx.recognition.OnnxRecognitionPredictor

All Implemented Interfaces:

IPredictor<BufferedImage,String>, IRecognitionPredictor, AutoCloseable

       public class OnnxRecognitionPredictor extends AbstractOnnxPredictor<BufferedImage,String> implements IRecognitionPredictor
      

A text recognition predictor implementation, which is using ONNX Runtime and its ML models to recognize text characters on an image.

Constructor Summary

Constructors

Constructor	Description
OnnxRecognitionPredictor(OnnxRecognitionPredictorProperties properties)	Creates a text recognition predictor with the specified properties.

Method Summary

Modifier and Type	Method	Description
static OnnxRecognitionPredictor	crnnMobileNetV3(String modelPath)	Creates a new text recognition predictor using an existing pre-trained CRNN model with a MobileNet V3 backbone, stored on disk.
static OnnxRecognitionPredictor	crnnMobileNetV3(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained CRNN model with a MobileNet V3 backbone, stored on disk.
static OnnxRecognitionPredictor	crnnVgg16(String modelPath)	Creates a new text recognition predictor using an existing pre-trained CRNN model with a VGG-16 backbone, stored on disk.
static OnnxRecognitionPredictor	crnnVgg16(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained CRNN model with a VGG-16 backbone, stored on disk.
static OnnxRecognitionPredictor	easyOcr(String modelPath, EasyOcrMapper labelMapper)	Creates a new text recognition predictor using an existing pre-trained EasyOCR model, stored on disk.
static OnnxRecognitionPredictor	easyOcr(String modelPath, EasyOcrMapper labelMapper, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained EasyOCR model, stored on disk.
protected List<String>	fromOutputBuffer(List<BufferedImage> inputBatch, FloatBufferMdArray outputBatch)	Converts ONNX runtime model batched output MD-array buffer to a list of predictor outputs.
OnnxRecognitionPredictorProperties	getProperties()	Returns the text recognition predictor properties.
static OnnxRecognitionPredictor	master(String modelPath)	Creates a new text recognition predictor using an existing pre-trained MASTER model, stored on disk.
static OnnxRecognitionPredictor	master(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained MASTER model, stored on disk.
static OnnxRecognitionPredictor	paddleOcr(String modelDirPath)	Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.
static OnnxRecognitionPredictor	paddleOcr(String modelDirPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.
static OnnxRecognitionPredictor	paddleOcr(String modelPath, String configPath)	Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.
static OnnxRecognitionPredictor	paddleOcr(String modelPath, String configPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.
static OnnxRecognitionPredictor	parSeq(String modelPath)	Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.
static OnnxRecognitionPredictor	parSeq(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.
static OnnxRecognitionPredictor	parSeq(String modelPath, Vocabulary vocabulary, int additionalTokens)	Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.
static OnnxRecognitionPredictor	parSeq(String modelPath, Vocabulary vocabulary, int additionalTokens, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.
Iterator<String>	predict(Iterator<BufferedImage> inputs)	Performs prediction on a sequence of input items.
static OnnxRecognitionPredictor	sar(String modelPath)	Creates a new text recognition predictor using an existing pre-trained SAR model, stored on disk.
static OnnxRecognitionPredictor	sar(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained SAR model, stored on disk.
protected FloatBufferMdArray	toInputBuffer(List<BufferedImage> batch)	Converts predictor inputs to an ONNX runtime model batched input MD-array buffer.
static OnnxRecognitionPredictor	viTstr(String modelPath)	Creates a new text recognition predictor using an existing pre-trained ViTSTR model, stored on disk.
static OnnxRecognitionPredictor	viTstr(String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)	Creates a new text recognition predictor using an existing pre-trained ViTSTR model, stored on disk.

Methods inherited from class com.itextpdf.pdfocr.onnx.AbstractOnnxPredictor

close

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface java.lang.AutoCloseable

close

Methods inherited from interface com.itextpdf.pdfocr.onnx.IPredictor

predict

Constructor Details

OnnxRecognitionPredictor

             public OnnxRecognitionPredictor
             (OnnxRecognitionPredictorProperties properties)
            

Creates a text recognition predictor with the specified properties.

Parameters:

properties - properties of the predictor

Method Details

crnnVgg16

             public static OnnxRecognitionPredictor crnnVgg16
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained CRNN model with a VGG-16 backbone, stored on disk. This is the default text recognition model in OnnxTR.

This can be used to load the following models from OnnxTR:

• crnn_vgg16_bn
• crnn_vgg16_bn (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the CRNN model loaded with a VGG-16 backbone

crnnVgg16

             public static OnnxRecognitionPredictor crnnVgg16
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained CRNN model with a VGG-16 backbone, stored on disk. This is the default text recognition model in OnnxTR.

This can be used to load the following models from OnnxTR:

• crnn_vgg16_bn
• crnn_vgg16_bn (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the CRNN model loaded with a VGG-16 backbone

crnnMobileNetV3

             public static OnnxRecognitionPredictor crnnMobileNetV3
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained CRNN model with a MobileNet V3 backbone, stored on disk.

This can be used to load the following models from OnnxTR:

• crnn_mobilenet_v3_large
• crnn_mobilenet_v3_large (8-bit quantized)
• crnn_mobilenet_v3_small
• crnn_mobilenet_v3_small (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the CRNN model loaded with a MobileNet V3 backbone

crnnMobileNetV3

             public static OnnxRecognitionPredictor crnnMobileNetV3
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained CRNN model with a MobileNet V3 backbone, stored on disk.

This can be used to load the following models from OnnxTR:

• crnn_mobilenet_v3_large
• crnn_mobilenet_v3_large (8-bit quantized)
• crnn_mobilenet_v3_small
• crnn_mobilenet_v3_small (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the CRNN model loaded with a MobileNet V3 backbone

master

             public static OnnxRecognitionPredictor master
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained MASTER model, stored on disk.

This can be used to load the following models from OnnxTR:

• MASTER
• MASTER (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the MASTER model loaded

master

             public static OnnxRecognitionPredictor master
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained MASTER model, stored on disk.

This can be used to load the following models from OnnxTR:

• MASTER
• MASTER (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the MASTER model loaded

parSeq

             public static OnnxRecognitionPredictor parSeq
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.

This can be used to load the following models from OnnxTR:

• parseq
• parseq (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the PARSeq model loaded

parSeq

             public static OnnxRecognitionPredictor parSeq
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.

This can be used to load the following models from OnnxTR:

• parseq
• parseq (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the PARSeq model loaded

parSeq

             public static OnnxRecognitionPredictor parSeq
             (String modelPath, Vocabulary vocabulary, int additionalTokens)
            

Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.

This can be used to load the following models from OnnxTR:

• parseq
• parseq (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

vocabulary - vocabulary used for the model output (without special tokens)

additionalTokens - amount of additional tokens in the total vocabulary after the end-of-string token

Returns:

a new predictor object with the PARSeq model loaded

parSeq

             public static OnnxRecognitionPredictor parSeq
             (String modelPath, Vocabulary vocabulary, int additionalTokens, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained PARSeq model, stored on disk.

This can be used to load the following models from OnnxTR:

• parseq
• parseq (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

vocabulary - vocabulary used for the model output (without special tokens)

additionalTokens - amount of additional tokens in the total vocabulary after the end-of-string token

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the PARSeq model loaded

sar

             public static OnnxRecognitionPredictor sar
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained SAR model, stored on disk.

This can be used to load the following models from OnnxTR:

• sar_resnet31
• sar_resnet31 (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the SAR model loaded

sar

             public static OnnxRecognitionPredictor sar
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained SAR model, stored on disk.

This can be used to load the following models from OnnxTR:

• sar_resnet31
• sar_resnet31 (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the SAR model loaded

viTstr

             public static OnnxRecognitionPredictor viTstr
             (String modelPath)
            

Creates a new text recognition predictor using an existing pre-trained ViTSTR model, stored on disk.

This can be used to load the following models from OnnxTR:

• vitstr_base
• vitstr_base (8-bit quantized)
• vitstr_small
• vitstr_small (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

Returns:

a new predictor object with the ViTSTR model loaded

viTstr

             public static OnnxRecognitionPredictor viTstr
             (String modelPath, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained ViTSTR model, stored on disk.

This can be used to load the following models from OnnxTR:

• vitstr_base
• vitstr_base (8-bit quantized)
• vitstr_small
• vitstr_small (8-bit quantized)

These models cannot handle spaces. Make sure you choose a detection model that outputs words.

Parameters:

modelPath - path to the pre-trained model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the ViTSTR model loaded

paddleOcr

             public static OnnxRecognitionPredictor paddleOcr
             (String modelDirPath) throws IOException
            

Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, PaddleOCR does not provide models in the ONNX format, you might need to do a model conversion yourself. Check out this page for information on how to do that.

This method expects the directory to contain two files:

• inference.onnx - the inference model in the ONNX format
• inference.yml - the configuration file for the model in YAML

This method can be used to load the following PaddleOCR models:

• PP-OCRv5_server_rec
• PP-OCRv5_mobile_rec
• PP-OCRv4_server_rec_doc
• PP-OCRv4_mobile_rec
• PP-OCRv4_server_rec
• PP-OCRv3_mobile_rec
• ch_SVTRv2_rec
• ch_RepSVTR_rec
• en_PP-OCRv5_mobile_rec
• en_PP-OCRv4_mobile_rec
• en_PP-OCRv3_mobile_rec
• korean_PP-OCRv5_mobile_rec
• latin_PP-OCRv5_mobile_rec
• eslav_PP-OCRv5_mobile_rec
• th_PP-OCRv5_mobile_rec
• el_PP-OCRv5_mobile_rec
• arabic_PP-OCRv5_mobile_rec
• cyrillic_PP-OCRv5_mobile_rec
• devanagari_PP-OCRv5_mobile_rec
• te_PP-OCRv5_mobile_rec
• ta_PP-OCRv5_mobile_rec
• korean_PP-OCRv3_mobile_rec
• japan_PP-OCRv3_mobile_rec
• chinese_cht_PP-OCRv3_mobile_rec
• te_PP-OCRv3_mobile_rec
• ka_PP-OCRv3_mobile_rec
• ta_PP-OCRv3_mobile_rec
• latin_PP-OCRv3_mobile_rec
• arabic_PP-OCRv3_mobile_rec
• cyrillic_PP-OCRv3_mobile_rec
• devanagari_PP-OCRv3_mobile_rec

These models can handle spaces.

Parameters:

modelDirPath - path to the directory with the model and its configuration file

Returns:

a new predictor object with the PaddleOCR model loaded

Throws:

IOException - if any I/O error occurs while loading configuration file

paddleOcr

             public static OnnxRecognitionPredictor paddleOcr
             (String modelDirPath, IOrtSessionOptionsCreator ortSessionOptionsCreator) throws IOException
            

Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, PaddleOCR does not provide models in the ONNX format, you might need to do a model conversion yourself. Check out this page for information on how to do that.

This method expects the directory to contain two files:

• inference.onnx - the inference model in the ONNX format
• inference.yml - the configuration file for the model in YAML

This method can be used to load the following PaddleOCR models:

• PP-OCRv5_server_rec
• PP-OCRv5_mobile_rec
• PP-OCRv4_server_rec_doc
• PP-OCRv4_mobile_rec
• PP-OCRv4_server_rec
• PP-OCRv3_mobile_rec
• ch_SVTRv2_rec
• ch_RepSVTR_rec
• en_PP-OCRv5_mobile_rec
• en_PP-OCRv4_mobile_rec
• en_PP-OCRv3_mobile_rec
• korean_PP-OCRv5_mobile_rec
• latin_PP-OCRv5_mobile_rec
• eslav_PP-OCRv5_mobile_rec
• th_PP-OCRv5_mobile_rec
• el_PP-OCRv5_mobile_rec
• arabic_PP-OCRv5_mobile_rec
• cyrillic_PP-OCRv5_mobile_rec
• devanagari_PP-OCRv5_mobile_rec
• te_PP-OCRv5_mobile_rec
• ta_PP-OCRv5_mobile_rec
• korean_PP-OCRv3_mobile_rec
• japan_PP-OCRv3_mobile_rec
• chinese_cht_PP-OCRv3_mobile_rec
• te_PP-OCRv3_mobile_rec
• ka_PP-OCRv3_mobile_rec
• ta_PP-OCRv3_mobile_rec
• latin_PP-OCRv3_mobile_rec
• arabic_PP-OCRv3_mobile_rec
• cyrillic_PP-OCRv3_mobile_rec
• devanagari_PP-OCRv3_mobile_rec

These models can handle spaces.

Parameters:

modelDirPath - path to the directory with the model and its configuration file

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the PaddleOCR model loaded

Throws:

IOException - if any I/O error occurs while loading configuration file

paddleOcr

             public static OnnxRecognitionPredictor paddleOcr
             (String modelPath, String configPath) throws IOException
            

Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, PaddleOCR does not provide models in the ONNX format, you might need to do a model conversion yourself. Check out this page for information on how to do that.

This method can be used to load the following PaddleOCR models:

• PP-OCRv5_server_rec
• PP-OCRv5_mobile_rec
• PP-OCRv4_server_rec_doc
• PP-OCRv4_mobile_rec
• PP-OCRv4_server_rec
• PP-OCRv3_mobile_rec
• ch_SVTRv2_rec
• ch_RepSVTR_rec
• en_PP-OCRv5_mobile_rec
• en_PP-OCRv4_mobile_rec
• en_PP-OCRv3_mobile_rec
• korean_PP-OCRv5_mobile_rec
• latin_PP-OCRv5_mobile_rec
• eslav_PP-OCRv5_mobile_rec
• th_PP-OCRv5_mobile_rec
• el_PP-OCRv5_mobile_rec
• arabic_PP-OCRv5_mobile_rec
• cyrillic_PP-OCRv5_mobile_rec
• devanagari_PP-OCRv5_mobile_rec
• te_PP-OCRv5_mobile_rec
• ta_PP-OCRv5_mobile_rec
• korean_PP-OCRv3_mobile_rec
• japan_PP-OCRv3_mobile_rec
• chinese_cht_PP-OCRv3_mobile_rec
• te_PP-OCRv3_mobile_rec
• ka_PP-OCRv3_mobile_rec
• ta_PP-OCRv3_mobile_rec
• latin_PP-OCRv3_mobile_rec
• arabic_PP-OCRv3_mobile_rec
• cyrillic_PP-OCRv3_mobile_rec
• devanagari_PP-OCRv3_mobile_rec

These models can handle spaces.

Parameters:

modelPath - path to the pre-trained model in the ONNX format

configPath - path to the configuration file for the model

Returns:

a new predictor object with the PaddleOCR model loaded

Throws:

IOException - if any I/O error occurs while loading configuration file

paddleOcr

             public static OnnxRecognitionPredictor paddleOcr
             (String modelPath, String configPath, IOrtSessionOptionsCreator ortSessionOptionsCreator) throws IOException
            

Creates a new text recognition predictor using an existing pre-trained PaddleOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, PaddleOCR does not provide models in the ONNX format, you might need to do a model conversion yourself. Check out this page for information on how to do that.

This method can be used to load the following PaddleOCR models:

• PP-OCRv5_server_rec
• PP-OCRv5_mobile_rec
• PP-OCRv4_server_rec_doc
• PP-OCRv4_mobile_rec
• PP-OCRv4_server_rec
• PP-OCRv3_mobile_rec
• ch_SVTRv2_rec
• ch_RepSVTR_rec
• en_PP-OCRv5_mobile_rec
• en_PP-OCRv4_mobile_rec
• en_PP-OCRv3_mobile_rec
• korean_PP-OCRv5_mobile_rec
• latin_PP-OCRv5_mobile_rec
• eslav_PP-OCRv5_mobile_rec
• th_PP-OCRv5_mobile_rec
• el_PP-OCRv5_mobile_rec
• arabic_PP-OCRv5_mobile_rec
• cyrillic_PP-OCRv5_mobile_rec
• devanagari_PP-OCRv5_mobile_rec
• te_PP-OCRv5_mobile_rec
• ta_PP-OCRv5_mobile_rec
• korean_PP-OCRv3_mobile_rec
• japan_PP-OCRv3_mobile_rec
• chinese_cht_PP-OCRv3_mobile_rec
• te_PP-OCRv3_mobile_rec
• ka_PP-OCRv3_mobile_rec
• ta_PP-OCRv3_mobile_rec
• latin_PP-OCRv3_mobile_rec
• arabic_PP-OCRv3_mobile_rec
• cyrillic_PP-OCRv3_mobile_rec
• devanagari_PP-OCRv3_mobile_rec

These models can handle spaces.

Parameters:

modelPath - path to the pre-trained model in the ONNX format

configPath - path to the configuration file for the model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the PaddleOCR model loaded

Throws:

IOException - if any I/O error occurs while loading configuration file

easyOcr

             public static OnnxRecognitionPredictor easyOcr
             (String modelPath, EasyOcrMapper labelMapper)
            

Creates a new text recognition predictor using an existing pre-trained EasyOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, EasyOCR does not provide models in the ONNX format, you might need to do a model conversion yourself.

This method can be used to load the following EasyOCR models:

• english_g2
• latin_g2
• zh_sim_g2
• japanese_g2
• korean_g2
• telugu_g2
• kannada_g2
• latin_g1
• zh_sim_g1
• zh_tra_g1
• japanese_g1
• korean_g1
• thai_g1
• devanagari_g1
• cyrillic_g1
• arabic_g1
• bengali_g1

These models can handle spaces.

Parameters:

modelPath - path to the pre-trained model in the ONNX format

labelMapper - label mapper to use for the model

Returns:

a new predictor object with the EasyOCR model loaded

easyOcr

             public static OnnxRecognitionPredictor easyOcr
             (String modelPath, EasyOcrMapper labelMapper, IOrtSessionOptionsCreator ortSessionOptionsCreator)
            

Creates a new text recognition predictor using an existing pre-trained EasyOCR model, stored on disk.

Only models in the ONNX format are supported. Since, by default, EasyOCR does not provide models in the ONNX format, you might need to do a model conversion yourself.

This method can be used to load the following EasyOCR models:

• english_g2
• latin_g2
• zh_sim_g2
• japanese_g2
• korean_g2
• telugu_g2
• kannada_g2
• latin_g1
• zh_sim_g1
• zh_tra_g1
• japanese_g1
• korean_g1
• thai_g1
• devanagari_g1
• cyrillic_g1
• arabic_g1
• bengali_g1

These models can handle spaces.

Parameters:

modelPath - path to the pre-trained model in the ONNX format

labelMapper - label mapper to use for the model

ortSessionOptionsCreator - the ONNX runtime session options creator

Returns:

a new predictor object with the EasyOCR model loaded

getProperties

             public OnnxRecognitionPredictorProperties getProperties()
            

Returns the text recognition predictor properties.

Returns:

the text recognition predictor properties

predict

             public Iterator<String> predict
             (Iterator<BufferedImage> inputs)
            

Performs prediction on a sequence of input items.

This method consumes the provided Iterator of inputs and produces an Iterator of outputs, typically yielding one result per input item.

Specified by:

predict in interface IPredictor<BufferedImage,String>

Overrides:

predict in class AbstractOnnxPredictor<BufferedImage,String>

Parameters:

inputs - an Iterator over the input items to be processed

Returns:

an Iterator over the predicted output items

toInputBuffer

             protected FloatBufferMdArray toInputBuffer
             (List<BufferedImage> batch)
            

Converts predictor inputs to an ONNX runtime model batched input MD-array buffer.

Specified by:

toInputBuffer in class AbstractOnnxPredictor<BufferedImage,String>

Parameters:

batch - batch of raw predictor inputs

Returns:

batched model input MD-array buffer

fromOutputBuffer

             protected List<String> fromOutputBuffer
             (List<BufferedImage> inputBatch, FloatBufferMdArray outputBatch)
            

Converts ONNX runtime model batched output MD-array buffer to a list of predictor outputs.

Specified by:

fromOutputBuffer in class AbstractOnnxPredictor<BufferedImage,String>

Parameters:

inputBatch - list of raw predictor inputs, matching the output

outputBatch - batched model output MD-array buffer

Returns:

a list of predictor output