Word Spotting and Recognition with Embedded Attributes

By Jon Almazán, Albert Gordo, Alicia Fornés, Ernest Valveny. 2014

528 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Natural Language Processing

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

In Natural Language Processing

Citations in the corpus

TAP: Text-Aware Pre-training for Text-VQA and Text-Caption (2020, positive): “An offline OCR detector [10] detects scene text regions and represents the region with its visual, coordinates, FastText [9], and Pyramidal Histogram of Characters (PHOC) [2] features.”
Deep Structured Output Learning for Unconstrained Text Recognition (2014, methodology): “Almaz´an et al. (2014) and more recently Gordo (2014) also formluate joint embedding spaces, achieving impressive results with minimal training data.”
Synthetic Data and Artificial Neural Networks for Natural Scene Text Recognition (2014, neutral): “3 - Almazan [1] - 89.”
Reading Text in the Wild with Convolutional Neural Networks (2014, background): “2013), text recognition (Almazán et al. 2014;Bissacco et al. 2013; Jaderberg et al. 2014;Mishra et al. 2012; Novikova et al. 2012; Rath and Manmatha 2007; Yao et al. 2014), or on combining both in end-to-end systems (Alsharif and Pineau 2014; Gordo 2014; Jaderberg et al.”
Scene Text Visual Question Answering (2019, background): “Scene Text Retrieval: This baseline leverages a single shot CNN architecture that predicts at the same time bounding boxes and a compact text representation (PHOC [2]) of the words in them [8].”
Finding the Evidence: Localization-aware Answer Prediction for Text Visual Question Answering (2020, methodology): “The OCR tokens are represented by a concatenation of the appearance features from Faster R-CNN , FastText embeddings (Bojanowski et al., 2017) , PHOC feature (Almazán et al., 2014) and bounding box (bbox) embedding.”
TextProposals: A text-specific selective search algorithm for word spotting in the wild (2016, methodology): “Moreover, the combination of our object proposals with existing whole-word recognizers [2, 3] shows state-of-the-art endto-end word spotting performance in several standard datasets, and, in some benchmarks, outperforms previously published results with a noticeable gain.”
Symmetry-Constrained Rectification Network for Scene Text Recognition (2019, unknown): “0 - 76.0 62.0 - - - - - Mishra et al. [38] 64.1 57.5 - 73.2 - 81.8 67.8 - - - - - Wang et al. [51] - - - 70.0 - 90.0 84.0 - - - - - Bissacco et al. [5] - - - - - 90.4 78.0 - 87.6 - - - Almazan et al. [1] 91.2 82.1 - 89.2 - - - - - - - - Yao et al. [55] 80.2 69.3 - 75.9 - 88.5 80.3 - - - - - Rodr´ıgu”
Confidence-aware Non-repetitive Multimodal Transformers for TextCaps (2020, unknown): “2017), Faster R-CNN, PHOC (Almazán et al. 2014) to extract subword feature x, appearance feature x and character-level feature x respectively.”
Structured Multimodal Attentions for TextVQA (2020, methodology): “For text nodes, we also employ a combination of multiple features (referred to as Multi-Feats) to enrich OCR regions’ representation as in [16]: 1) a 300D FastText feature {xft,i} N+M i=N+1 is generated from a pre-trained FastText [18] embeddings, 2) a 2048D appearance feature {xfr,i} N+M i=N+1 is g”
ESIR: End-To-End Scene Text Recognition via Iterative Image Rectification (2018, unknown): “CDAR2013 ICDAR2015 IIIT5K SVT SVTP CUTE None None 50 1k None 50 None None None Wang [44] [-] - - - - - 70.0 - - - Bissacco [4] [-] 87.6 - - - - - - - - Yao [47] [-] - - 80.2 69.3 - 75.9 - - - Almazan [1] [-] - 91.2 82.1 - 89.2 - - - Gordo [10] [-] - 93.3 86.6 - 91.8 - - - Jaderberg [16] [VGG, SK] 81”
Integrating Scene Text and Visual Appearance for Fine-Grained Image Classification (2017, methodology): “In the recognition stage, character-based methods [31], [32], [48]–[50] and word-based methods [6], [37], [51] are the two mainstream recognition approaches.”
An End-to-End Trainable Neural Network for Image-Based Sequence Recognition and Its Application to Scene Text Recognition (2015, background): “Notice that though the recent models learned by label embedding [11], [15] and incre-”
Robust Scene Text Recognition with Automatic Rectification (2016, background): “[2] propose to predict label embedding vectors from input images.”
ASTER: An Attentional Scene Text Recognizer with Flexible Rectification (2019, unknown): “[2] 91.”
Multimodal Attention with Image Text Spatial Relationship for OCR-Based Image Captioning (2020, methodology): “For OCR tokens, we also consider the semantic features which can be obtained from the pretrained FastText [7] vectors and Pyramidal Histogram of Characters [1] vectors.”
Vectorization and Rasterization: Self-Supervised Learning for Sketch and Handwriting (2021, methodology): “Following this seminal work [20], earlier handcrafted feature extraction methods [3, 26, 1] in both the modalities have now been replaced by data driven feature learning [5, 50, 9] as in many computer vision domains.”
Recursive Recurrent Nets with Attention Modeling for OCR in the Wild (2016, unknown): “[1] 89.”