Scene Text Localization and Recognition with Oriented Stroke Detection

By Lukás Neumann, Jiri Matas. 2013

231 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology, Information Extraction, Human-Computer Interaction

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. Chunhua Shen
3. Luc Van Gool
4. Dacheng Tao
5. Yu Qiao
6. Xiaogang Wang
7. Andrew Zisserman
8. Cong Yao

In Media Technology

1. Cong Yao
2. Xiang Bai
3. Chunhua Shen
4. Yu Qiao
5. Andrea Vedaldi
6. Chen Change Loy
7. Tong He
8. Weilin Huang

In Information Extraction

1. Jian Yao
2. Feng Su
3. Yu Qiao
4. Peng Wang
5. Xiang Bai
6. Andrea Vedaldi
7. Andrew Zisserman
8. Qixiang Ye

In Human-Computer Interaction

1. Xiaoou Tang
2. Cong Yao
3. Xiang Bai
4. Andrew Zisserman
5. Karen Simonyan
6. Jian Yao
7. Feng Su
8. Yu Qiao

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].”
• Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition (2014, background): “More recent works such as [2, 3, 20] make use of over-segmentation methods, guided by a supervised classifier, to generate candidate proposals which are subsequently classified as characters or false positives.”
• A Unified Framework for Multioriented Text Detection and Recognition (2014, methodology): “For case sensitive recognition, our system achieves a performance of precision 0.492, recall 0.440 and F-measure 0.454, which slightly outperforms the system of Neumann et al. [75] but is inferior than the method of Milyaev et al. [74].”
• Text Detection and Recognition in Imagery: A Survey (2015, methodology): “Neumann and Matas [189] introduced an end-to-end”
• Deep Features for Text Spotting (2014, methodology): “Text detection methods are either based on connected components (CCs) [11, 16, 50, 49, 32–35] or sliding windows [40, 7, 39, 45].”
• Symmetry-based text line detection in natural scenes (2015, methodology): “There are mainly two classes of mainstream methods for scene text detection: those based on a sliding window [5, 36, 27] and those based on connected component extraction [7, 25, 9].”
• Arbitrary-Oriented Scene Text Detection via Rotation Proposals (2017, methodology): “Sliding window-based methods [7], [10], [28]–[30] tend to use a sliding window of a fixed size to slide the text area and find the region most likely to include text.”
• Robust Seed Localization and Growing with Deep Convolutional Features for Scene Text Detection (2015, methodology): “Neumann et al. [13] combines the sliding-window and connected component methods.”
• Reading Scene Text in Deep Convolutional Sequences (2015, background): “This problem includes two sub tasks, namely text detection [16, 34, 35, 15, 24] and text-line/word recognition[18, 1, 17, 4, 33].”
• Multi-oriented Text Detection with Fully Convolutional Networks (2016, methodology): “sliding windows [8, 25, 3, 24, 17] and connected component [16, 6, 4, 32, 28] have become mainstream in this specific domain.”
• Object Detection in 20 Years: A Survey (2019, methodology): “emal Regions (MSER) segmentation [341] and morphological filtering [346]. Some domain knowledge, such as the symmetry of texts and the structures of strokes, also have been considered in these methods [340, 341, 347]. In recent years, researchers have paid more attention to the problem of text locali”
• TextProposals: A text-specific selective search algorithm for word spotting in the wild (2016, methodology): “Another method that combines the advantages of sliding-window and region-based approaches is proposed by Neumann and Matas in [9], where characters are detected as image regions that contain certain strokes with specific orientations in specific positions.”
• Scene text detection and recognition: recent advances and future trends (2015, background): “[66] presented a new system for scene text localization and recognition, which combines the advantages of sliding-window based and component based methods.”
• Robust Scene Text Detection with Convolution Neural Network Induced MSER Trees (2014, methodology): “In contrast to previous works, which often use a set of heuristic features to distinguish text and non-text components [23,24,22,34,31,10], we take the advantages of deep learning and adapt”