Scene Text Recognition from Two-Dimensional Perspective

Abstract

Inspired by speech recognition, recent state-of-the-art algorithms mostly consider scene text recognition as a sequence prediction problem. Though achieving excellent performance, these methods usually neglect an important fact that text in images are actually distributed in two-dimensional space. It is a nature quite different from that of speech, which is essentially a one-dimensional signal. In principle, directly compressing features of text into a one-dimensional form may lose useful information and introduce extra noise. In this paper, we approach scene text recognition from a two-dimensional perspective. A simple yet effective model, called Character Attention Fully Convolutional Network (CA-FCN), is devised for recognizing the text of arbitrary shapes. Scene text recognition is realized with a semantic segmentation network, where an attention mechanism for characters is adopted. Combined with a word formation module, CA-FCN can simultaneously recognize the script and predict the position of each character. Experiments demonstrate that the proposed algorithm outperforms previous methods on both regular and irregular text datasets. Moreover, it is proven to be more robust to imprecise localizations in the text detection phase, which are very common in practice.

FACLSTM: ConvLSTM with focused attention for scene text recognition

“…Different from those LSTM-based approaches, recognizers without LSTM can better leverage the spatial information, but they also unavoidably introduce additional parameters or post processing steps in order to produce sequential outputs, such as the multiple classifiers used by STN-OCR [9] and the word formation module designed in [10].…”

“…Backbone: Similar to Liao's work [10], we take VGG-16 as the encoder of our feature extraction module, and remove the fully connected layers and pooling layers from the last two encoding stages. We also assemble two deformable convolutional layers [24] at stage-4 and stage-5 of the decoder given their flexible receptive fields.…”

“…We also assemble two deformable convolutional layers [24] at stage-4 and stage-5 of the decoder given their flexible receptive fields. However, compared with Liao's network [10], the resolution of final feature maps is restored to a smaller size of W 4 × H 4 × C in our FACLSTM, instead of the W 2 × H 2 × C used in [10], considering the memory and computation cost. Here, W , H and C denote the width, height and channels of feature maps, respectively.…”

A New Perspective for Flexible Feature Gathering in Scene Text Recognition Via Character Anchor Pooling

“…Recently, CA-FCN [11] takes the two-dimensional spatial distribution of text into consideration, and text recognition is reformulated as semantic segmentation, where character categories are segmented from the background. However, their method abandons the use of recurrent neural networks (RNN), and thus fails to obtain an overall vision.…”

Recognition of images of Korean characters using embedded networks

“…Over the years, optical character recognition has been a popular research topic for computer vision specialists [1][2][3][4][5][6]. Convolutional neural networks have proven themselves as a good solution for such problems as object recognition.…”