Real-time scene text localization and recognition

By Lukás Neumann, Jiri Matas. 2012

879 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Xiang Bai
2. Cong Yao
3. Chunhua Shen
4. Luc Van Gool
5. Yu Qiao
6. Andrew Zisserman
7. Andrea Vedaldi
8. Dacheng Tao

Citations in the corpus

• Speeding up Convolutional Neural Networks with Low Rank Expansions (2014, background): “Character classification is an essential part of many text spotting pipelines such as [3, 4, 15, 24, 25, 26, 27, 29, 30, 41, 43].”
• TGIF: A New Dataset and Benchmark on Animated GIF Description (2016, methodology): “We filter out GIFs that contain text, e.g., memes, by detecting text regions using the Extremal Regions detector [23] and discarding a GIF if the regions cover more than 2% of the image area.”
• Synthetic Data for Text Localisation in Natural Images (2016, background): “[33, 34], while the recall is significantly worse by 12%.”
• Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition (2014, methodology): “Previous work uses sliding window methods [26] or region grouping methods [5, 6, 19] very successfully for this.”
• COCO-Text: Dataset and Benchmark for Text Detection and Recognition in Natural Images (2016, background): “Further, Neumann and Matas [14] first identify Ex-tremal Regions, groups them into words and then selects most probable character segmentation.”
• Deep Features for Text Spotting (2014, background): “Authors have subsequently focused solely on text detection [7, 11, 16, 50, 51], or text recognition [31, 36, 41], or on combining both in end-to-end systems [40, 39, 49, 32–34, 45, 35, 6, 8, 48].”
• Image Retrieval Using Textual Cues (2013, methodology): “The querydriven approach we propose outperforms localization and recognition pipeline based methods [9, 15].”
• Scene text detection and recognition: recent advances and future trends (2015, background): “11 End-to-end recognition examples of [13] (image reprinted from Ref.”
• Single Shot Text Detector with Regional Attention (2017, methodology): “, stroke width [3, 9] or extremal region [22, 10, 33]), or sliding window methods [12, 30, 6].”
• TextProposals: A text-specific selective search algorithm for word spotting in the wild (2016, methodology): “model [18, 26, 33, 25], or because they learn an optimal grouping strategy from examples of well-segmented character groupings [29, 34].”
• Decoupled Attention Network for Text Recognition (2019, background): “Early work of scene text recognition relied on low-level features, such as histogram of oriented gradients descriptors (Wang, Babenko, and Belongie 2011), connected components (Neumann and Matas 2012), etc.”
• ESIR: End-To-End Scene Text Recognition via Iterative Image Rectification (2018, methodology): “Different scene character detection and recognition methods have been reported by using sliding window [43, 42], connected components [31], extremal regions [32], Hough voting [47], co-occurrence histograms [41], etc.”
• TextBoxes++: A Single-Shot Oriented Scene Text Detector (2018, methodology): “targets: Most text detection methods can be roughly categorized into three categories: 1) Character-based: Individual characters or parts of the text are first detected and then grouped into words [7].”
• Real-Time Scene Text Detection With Differentiable Binarization and Adaptive Scale Fusion (2022, background): “Neumann and Matas [41] proposed to locate characters by classifying Extremal Regions (ERs).”
• Robust Seed Localization and Growing with Deep Convolutional Features for Scene Text Detection (2015, methodology): “Many existing methods are devoted to the first stage, which can be further categorized into two groups: the exhaustive search based methods [1, 10, 16, 18] and the selective search based methods [2, 12, 13, 6, 14, 21, 3, 9].”
• Focusing Attention: Towards Accurate Text Recognition in Natural Images (2017, background): “Scene text recognition has attracted much research interest of the computer vision community [11, 21, 24].”
• Scene Text Localization and Recognition with Oriented Stroke Detection (2013, methodology): “The problems of the method include ambiguities introduced by the fact that a subregion of a character might be another character, failures to detect letters on word boundaries which consist of just one stroke (e.g. “I”, “l”) and false positives caused by strokes around areas with text (see Figure 11”