태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
See Scene Text Detection for leaderboards in this task.
출처에 같은 태스크 이름으로 연결된 데이터셋입니다. 동명 태스크가 포함될 수 있으므로, 비교할 때는 평가 결과의 조건도 확인하세요.
The ICDAR 2013 dataset consists of 229 training images and 233 testing images, with word-level annotations provided. It is the standard benchmark dataset for evaluating near-horizontal text detection. Source: Single Shot Text Detector with Regional Attention Image Source: https://plos.figshare.com/articles/Detection_examples_of_the_proposed_method_on_the_ICDAR_2013_dataset_17_/5325856
The COCO-Text dataset is a dataset for text detection and recognition. It is based on the MS COCO dataset, which contains images of complex everyday scenes. The COCO-Text dataset contains non-text images, legible text images and illegible text images. In total there are 22184 training images and 7026 validation images with at least one instance of legible text. Source: Improving Text Proposals for Scene Images with Fully Convolutional Networks Image Source: https://vision.cornell.edu/se3/coco-text-2/
The ICDAR2003 dataset is a dataset for scene text recognition. It contains 507 natural scene images (including 258 training images and 249 test images) in total. The images are annotated at character level. Characters and words can be cropped from the images. Source: Robust Scene Text Recognition Using Sparse Coding based Features Image Source: https://www.researchgate.net/figure/The-results-of-text-localization-and-extraction-on-ICDAR-2003-dataset_fig3_290070044
TextOCR is a dataset to benchmark text recognition on arbitrary shaped scene-text. TextOCR requires models to perform text-recognition on arbitrary shaped scene-text present on natural images. TextOCR provides ~1M high quality word annotations on TextVQA images allowing application of end-to-end reasoning on downstream tasks such as visual question answering or image captioning. Dataset statistics: - 28,134 natural images from TextVQA - 903,069 annotated scene-text words - 32 words per image on average
The Street View Text (SVT) dataset was harvested from Google Street View. Image text in this data exhibits high variability and often has low resolution. In dealing with outdoor street level imagery, we note two characteristics. (1) Image text often comes from business signage and (2) business names are easily available through geographic business searches. These factors make the SVT set uniquely suited for word spotting in the wild: given a street view image, the goal is to identify words from nearby businesses. Note: the dataset has undergone revision since the time it was evaluated in this publication. Please consult the ICCV2011 paper for most up-to-date results. Source: Street View Text Dataset Image source: Street View Text Dataset
Features a large-scale dataset with 12,263 annotated images. Two tasks, namely text localization and end-to-end recognition, are set up. The competition took place from January 20 to May 31, 2017. 23 valid submissions were received from 19 teams. Source: ICDAR2017 Competition on Reading Chinese Text in the Wild (RCTW-17)
The heavily occluded scene text (HOST) dataset is a dataset that contains images of text with occlusions. It is used to improve the recognition performance of occluded text in machine vision applications 1. The dataset is composed of 4832 images that are manually occluded in weak or heavy degrees.
The CUTE80 dataset is a lightweight collection of images specifically designed for text detection in natural scene images. It contains a total of 13,000 annotated page images across five different popular categories: 1) Table 2) Figure 3) Natural image 4) Logo 5) ignature
The Weakly Occluded Scene Text (WOST) dataset is a public dataset for scene text segmentation. It is used to generate pixel-level annotations in scene text images 1. The dataset is designed to contain weakly annotated images, which means that the images are not fully annotated with pixel-level labels.
This dataset includes 4,500 fully annotated images (over 30,000 license plate characters) from 150 vehicles in real-world scenarios where both the vehicle and the camera (inside another vehicle) are moving. The images were acquired with three different cameras and are available in the Portable Network Graphics (PNG) format with a size of 1,920 × 1,080 pixels. The cameras used were: GoPro Hero4 Silver, Huawei P9 Lite, and iPhone 7 Plus. We collected 1,500 images with each camera, divided as follows: - 900 of cars with gray license plates; - 300 of cars with red license plates; - 300 of motorcycles with gray license plates. The dataset is split as follows: 40% for training, 40% for testing and 20% for validation. Every image has the following annotations available in a text file: the camera in which the image was taken, the vehicle’s position and information such as type (car or motorcycle), manufacturer, model and year; the identification and position of the license plate, as well as the position of its characters. Source: A Robust Real-Time Automatic License Plate Recognition Based on the YOLO Detector
The IIIT5K dataset contains 5,000 text instance images: 2,000 for training and 3,000 for testing. It contains words from street scenes and from originally-digital images. Every image is associated with a 50 -word lexicon and a 1,000 -word lexicon.
SVTP dataset stands for Scene Text Recognition Datasets. It is a collection of 4 popular Latin/English scene text recognition datasets, namely IIIT5K, SVT, SVTP, and CUTE-80. These datasets only provide case-insensitive annotations and no punctuation marks.
This dataset, called RodoSol-ALPR dataset, contains 20,000 images captured by static cameras located at pay tolls owned by the Rodovia do Sol (RodoSol) concessionaire, which operates 67.5 kilometers of a highway (ES-060) in the Brazilian state of Espírito Santo. There are images of different types of vehicles (e.g., cars, motorcycles, buses and trucks), captured during the day and night, from distinct lanes, on clear and rainy days, and the distance from the vehicle to the camera varies slightly. All images have a resolution of 1,280 × 720 pixels. An important feature of the proposed dataset is that it has images of two different license plate (LP) layouts: Brazilian and Mercosur (to maintain consistency with existing works, we refer to “Brazilian” as the standard used in Brazil before the adoption of the Mercosur standard). Every image has the following information available in a text file: the vehicle’s type (car or motorcycle), the LP’s layout (Brazilian or Mercosul), its text (e.g., ABC-1234), and the position (x, y) of each of its four corners. We labeled the corners instead of just the LP bounding box to enable the training of methods that explore LP rectification, as well as the application of a wider range of data augmentation techniques. Regarding privacy concerns related to our dataset, we remark that in Brazil the LPs are related to the respective vehicles, i.e., no public information is available about the vehicle drivers/owners. Moreover, all human faces (e.g., drivers or RodoSol’s employees) were manually redacted (i.e., blurred) in each image.
This dataset aims at evaluating the License Plate Character Segmentation (LPCS) problem. The experimental results of the paper Benchmark for License Plate Character Segmentation were obtained using a dataset providing 101 on-track vehicles captured during the day. The video was recorded using a static camera in early 2015. The images of the dataset were acquired with a digital camera in Full-HD and are available in the Portable Network Graphics (PNG) format with 1920×1080 pixels each. The average size of each file is 4.08 Megabytes (a total of 8.60 Gigabytes for the entire dataset). In addition, since there are some approaches that track the car to utilize redundant information to improve the recognition results, we decided to make a dataset with multiples frames per car. In this dataset, there are, on average, 19.80 image frames per vehicle (with a standard deviation of 4.14). Source: Benchmark for License Plate Character Segmentation
IIIT-ILST is a dataset and benchmark for scene text recognition for three Indic scripts - Devanagari, Telugu and Malayalam. IIIT-ILST contains nearly 1000 real images per each script which are annotated for scene text bounding boxes and transcriptions.
MLe2 is a dataset for the evaluation of scene text end-to-end reading systems and all intermediate stages such as text detection, script identification and text recognition. The dataset contains a total of 711 scene images covering four different scripts (Latin, Chinese, Kannada, and Hangul). Source: A fine-grained approach to scene text script identification
The ChineseLP dataset contains 411 vehicle images (mostly of passenger cars) with Chinese license plates (LPs). It consists of 252 images captured by the authors and 159 images downloaded from the internet. The images present great variations in resolution (from 143 × 107 to 2048 × 1536 pixels), illumination and background.
Vehicle-Rear is a novel dataset for vehicle identification that contains more than three hours of high-resolution videos, with accurate information about the make, model, color and year of nearly 3,000 vehicles, in addition to the position and identification of their license plates.
The Mobile Turkish Scene Text (MTST 200) dataset consists of 200 indoor and outdoor Turkish scene text images. The images were collected with mobile phones and downsized to 576 × 1024 (portrait) or 1024 × 576 (landscape) pixels. The text lines are horizontal or near horizontal, some with slight in– and out-of-plane rotations.
This dataset contains 2,000 images taken from inside a warehouse of the Energy Company of Paraná (Copel), which directly serves more than 4 million consuming units in the Brazilian state of Paraná. The images were acquired with three different cameras and are available in the JPG format with a resolution between 2,340 × 4,160 and 3,120 × 4,160 pixels. The dataset is split into three sets: training (800 images), validation (400 images) and testing (800 images). Every image has the following annotations available in a text file: the camera in which the image was taken, the counter’s position (x,y,w,h) and reading, as well as the position of each digit. All counters of the dataset (regardless of meter type) have 5 digits, and thus 10,000 digits were manually annotated. Source: Convolutional Neural Networks for Automatic Meter Reading
This dataset contains 12,500 meter images acquired in the field by the employees of the Energy Company of Paraná (Copel), which directly serves more than 4 million consuming units, across 395 cities and 1,113 locations (i.e., districts, villages and settlements), located in the Brazilian state of Paraná. Copel-AMR is composed of images captured in unconstrained scenarios, which typically include blur (due to camera motion), dirt, scale variations, in-plane and out-of-plane rotations, reflections, shadows, and occlusions. In 2,500 images (i.e., 20% of the dataset), it is not even possible to perform the meter reading due to occlusions or faulty meters. The images have a resolution of 480×640 or 640×480 pixels, depending on the orientation in which they were taken. Considering that the meter is operational and that there are no occlusions, these resolutions are enough for the meter reading to be legible. The dataset was randomly split as follows: 5,000 images for training, 5,000 images for testing and 2,500 images for validation, following the split protocol (i.e., 40%/40%/20%) used in the UFPR-AMR dataset. For reproducibility purposes, the subsets generated are explicitly available along with the Copel-AMR dataset. For each image in our dataset, we manually labeled the meter reading, the position (x, y) of each of the four corners of the counter, and a bounding box (x, y, w, h) for each digit. Corner annotations – which can be converted to a bounding box – enable the counter to be rectified, while bounding boxes enable the training of object detectors. Source: Towards Image-based Automatic Meter Reading in Unconstrained Scenarios: A Robust and Efficient Approach
The IC13 dataset contains 561 images: 420 for training and 141 for testing. It inherits data from the IC03 dataset and extends it with new images. Similar to IC03 dataset, the IC13 dataset contains 1,015 cropped text instance images after removing the words with non-alphanumeric characters.
The UTRSet-Real dataset is a comprehensive, manually annotated dataset specifically curated for Printed Urdu OCR research. It contains over 11,000 printed text line images, each of which has been meticulously annotated. One of the standout features of this dataset is its remarkable diversity, which includes variations in fonts, text sizes, colours, orientations, lighting conditions, noises, styles, and backgrounds. This diversity closely mirrors real-world scenarios, making the dataset highly suitable for training and evaluating models that aim to excel in real-world Urdu text recognition tasks. The availability of the UTRSet-Real dataset addresses the scarcity of comprehensive real-world printed Urdu OCR datasets. By providing researchers with a valuable resource for developing and benchmarking Urdu OCR models, this dataset promotes standardized evaluation and reproducibility and fosters advancements in the field of Urdu OCR. Further, to complement the UTRSet-Real for training purposes, we also present UTRSet-Synth, a high-quality synthetic dataset closely resembling real-world representations of Urdu text. For more information and details about the UTRSet-Real & UTRSet-Synth datasets, please refer to the paper "UTRNet: High-Resolution Urdu Text Recognition In Printed Documents"
The UTRSet-Synth dataset is introduced as a complementary training resource to the UTRSet-Real Dataset, specifically designed to enhance the effectiveness of Urdu OCR models. It is a high-quality synthetic dataset comprising 20,000 lines that closely resemble real-world representations of Urdu text. To generate the dataset, a custom-designed synthetic data generation module which offers precise control over variations in crucial factors such as font, text size, colour, resolution, orientation, noise, style, and background, was employed. Moreover, the UTRSet-Synth dataset tackles the limitations observed in existing datasets. It addresses the challenge of standardizing fonts by incorporating over 130 diverse Urdu fonts, which were thoroughly refined to ensure consistent rendering schemes. It overcomes the scarcity of Arabic words, numerals, and Urdu digits by incorporating a significant number of samples representing these elements. Additionally, the dataset is enriched by randomly selecting words from a vocabulary of 100,000 words during the text generation process. As a result, UTRSet-Synth contains a total of 28,187 unique words, with an average word length of 7 characters. The availability of the UTRSet-Synth dataset, a synthetic dataset that closely emulates real-world variations, addresses the scarcity of comprehensive real-world printed Urdu OCR datasets. By providing researchers with a valuable resource for developing and benchmarking Urdu OCR models, this dataset promotes standardized evaluation, and reproducibility, and fosters advancements in the field of Urdu OCR. For more information and details about the UTRSet-Real & UTRSet-Synth datasets, please refer to the paper "UTRNet: High-Resolution Urdu Text Recognition In Printed Documents"
This repository contains datasets and baselines for benchmarking Chinese text recognition. Please see the corresponding paper for more details regarding the datasets, baselines, the empirical study, etc. Highlights 🌟 All datasets are transformed to lmdb format for convenient usage. 🌟 The experimental results of all baselines are available at link with format (index [pred] [gt]). 🌟 The code and trained weights of all baselines are available at link for direct use.
This dataset contains images and annotations for scene text detection and recognition. It is made up of two parts: (1) 1,175 images manually labeled with a total of 59,588 text instances at the line and word levels; and (2) 929 signboard images collected from the VinText, Total-Text, and ICDAR15 datasets. Each text instance in the first part of our dataset has a quadrilateral bounding box and a ground truth character sequence associated with it. In the second part, images are selected if they contain signboards. This portion of the dataset comprises 20,261 text instances at word levels. This brings the total text instances in our final dataset up to 79,814. Following the ICDAR15 standard, we annotated each image with all of the text instances, polygons, and content that were present. Manual annotations were done on each and every image.
The Turkish Scene Text Recognition (TS-TR) dataset was primarily developed to fill the gap in non-English text recognition resources, specifically addressing the unique challenges presented by the Turkish language, such as special characters and diacritics. This dataset mirrors real-world conditions with texts displayed in various fonts, sizes, orientations, and complex backgrounds from multiple urban and rural environments. Such diversity ensures the training of models that can generalize across different scenarios, including varying lighting conditions and complex visual layouts.
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.