mmocr/docs/datasets.md

Datasets Preparation

This page lists the datasets which are commonly used in text detection, text recognition and key information extraction, and their download links.

Text Detection

The structure of the text detection dataset directory is organized as follows.

├── ctw1500
│   ├── imgs
│   ├── instances_test.json
│   └── instances_training.json
├── icdar2015
│   ├── imgs
│   ├── instances_test.json
│   └── instances_training.json
├── icdar2017
│   ├── imgs
│   ├── instances_training.json
│   └── instances_val.json
├── synthtext
│   ├── imgs
│   ├── instances_training.json
│   ├── instances_training.txt
│   └── instances_training.lmdb

Dataset		Images			Annotation Files			Note
				training	validation	testing
CTW1500		link		instances_training.json	-	instances_test.json
ICDAR2015		link		instances_training.json	-	instances_test.json
ICDAR2017		link		instances_training.json	instances_val.json	instances_test.json
Synthtext		link		instances_training.json instances_training.txt	-

• For icdar2015:
  • Step1: Download ch4_training_images.zip and ch4_test_images.zip from this link
  • Step2: Download instances_training.json and instances_test.json
  • Step3:

  mkdir icdar2015 && cd icdar2015
  mv /path/to/instances_training.json .
  mv /path/to/instances_test.json .
  
  mkdir imgs && cd imgs
  ln -s /path/to/ch4_training_images training
  ln -s /path/to/ch4_test_images test
  

Text Recognition

The structure of the text recognition dataset directory is organized as follows.

├── mixture
│   ├── coco_text
│   │   ├── train_label.txt
│   │   ├── train_words
│   ├── icdar_2011
│   │   ├── training_label.txt
│   │   ├── Challenge1_Training_Task3_Images_GT
│   ├── icdar_2013
│   │   ├── train_label.txt
│   │   ├── test_label_1015.txt
│   │   ├── test_label_1095.txt
│   │   ├── Challenge2_Training_Task3_Images_GT
│   │   ├── Challenge2_Test_Task3_Images
│   ├── icdar_2015
│   │   ├── train_label.txt
│   │   ├── test_label.txt
│   │   ├── ch4_training_word_images_gt
│   │   ├── ch4_test_word_images_gt
│   ├── III5K
│   │   ├── train_label.txt
│   │   ├── test_label.txt
│   │   ├── train
│   │   ├── test
│   ├── ct80
│   │   ├── test_label.txt
│   │   ├── image
│   ├── svt
│   │   ├── test_label.txt
│   │   ├── image
│   ├── svtp
│   │   ├── test_label.txt
│   │   ├── image
│   ├── Synth90k
│   │   ├── shuffle_labels.txt
│   │   ├── mnt
│   ├── SynthText
│   │   ├── shuffle_labels.txt
│   │   ├── instances_train.txt
│   │   ├── synthtext
│   ├── SynthAdd
│   │   ├── label.txt
│   │   ├── SynthText_Add

Dataset		images	annotation file	annotation file	Note
			training	test
coco_text		link	train_label.txt	-
icdar_2011		link	train_label.txt	-
icdar_2013		link	train_label.txt	test_label_1015.txt
icdar_2015		link	train_label.txt	test_label.txt
IIIT5K		link	train_label.txt	test_label.txt
ct80		-	-	test_label.txt
svt		link	-	test_label.txt
svtp		-	-	test_label.txt
Synth90k		link	shuffle_labels.txt	-
SynthText		link	shuffle_labels.txt | instances_train.txt	-
SynthAdd		link	label.txt	-

• For icdar_2013:

  • Step1: Download Challenge2_Test_Task3_Images.zip and Challenge2_Training_Task3_Images_GT.zip from this link
  • Step2: Download test_label_1015.txt and train_label.txt

• For icdar_2015:

  • Step1: Download ch4_training_word_images_gt.zip and ch4_test_word_images_gt.zip from this link
  • Step2: Download train_label.txt and test_label.txt

• For IIIT5K:

  • Step1: Download IIIT5K-Word_V3.0.tar.gz from this link
  • Step2: Download train_label.txt and test_label.txt

• For svt:

  • Step1: Download svt.zip form this link
  • Step2: Download test_label.txt

• For ct80:

  • Step1: Download test_label.txt

• For svtp:

  • Step1: Download test_label.txt

• For coco_text:

  • Step1: Download from this link
  • Step2: Download train_label.txt

• For Syn90k:

  • Step1: Download mjsynth.tar.gz from this link
  • Step2: Download shuffle_labels.txt
  • Step3:

  mkdir Syn90k && cd Syn90k
  
  mv /path/to/mjsynth.tar.gz .
  
  tar -xzf mjsynth.tar.gz
  
  mv /path/to/shuffle_labels.txt .
  
  # create soft link
  cd /path/to/mmocr/data/mixture
  
  ln -s /path/to/Syn90k Syn90k
  

• For SynthText:

  • Step1: Download SynthText.zip from this link
  • Step2: Download shuffle_labels.txt
  • Step3: Download instances_train.txt
  • Step4:

  unzip SynthText.zip
  
  cd SynthText
  
  mv /path/to/shuffle_labels.txt .
  
  # create soft link
  cd /path/to/mmocr/data/mixture
  
  ln -s /path/to/SynthText SynthText
  

• For SynthAdd:

  • Step1: Download SynthText_Add.zip from this link
  • Step2: Download label.txt
  • Step3:

  mkdir SynthAdd && cd SynthAdd
  
  mv /path/to/SynthText_Add.zip .
  
  unzip SynthText_Add.zip
  
  mv /path/to/label.txt .
  
  # create soft link
  cd /path/to/mmocr/data/mixture
  
  ln -s /path/to/SynthAdd SynthAdd