Revisiting Classification Perspective on Scene Text Recognition

Abstract

The prevalent perspectives of scene text recognition are from sequence to sequence (seq2seq) and segmentation. In this paper, we propose a new perspective on scene text recognition, in which we model the scene text recognition as an image classification problem. Based on the image classification perspective, a scene text recognition model is proposed, which is named as CSTR.The CSTR model consists of a series of convolutional layers and a global average pooling layer at the end, followed by independent multi-class classification heads, each of which predicts the corresponding character of the word sequence in input image. The CSTR model is easy to train using parallel cross entropy losses.CSTR is as simple as image classification models like ResNet [15] which makes it easy to implement, and the fully convolutional neural network architecture makes it efficient to train and deploy. We demonstrate the effectiveness of the classification perspective on scene text recognition with thorough experiments. Futhermore, CSTR achieves nearly state-of-the-art performance on six public benchmarks including regular text, irregular text. The code will be available at https://github.com/Media-Smart/vedastr.

HyText – A Scene-Text Extraction Method for Video Retrieval

“…In order to find a suitable STR component for our pipeline, we compare the performance of several methods proposed over the last few years. Specifically, we compare CRNN [19], TPS-ResNet-BiLSTM-CTC (TRBC), TPS-ResNet-BiLSTM-Attn (TRBA), SARN [9], CSTR [2] and the recognition module of EasyOCR (based on CRNN). All these methods except for the EasyOCR recognition module are trained on Synthtext [6] to allow for a fair comparison.…”

Attention Guided Feature Encoding for Scene Text Recognition

“…The PhotoOCR application in [ 12 ] demonstrated the use of deep neural networks without convolutional operations for character recognition with raw and edge-based feature representations. The recent work by Cai et al [ 14 ] explored the image classification methodology used for scene text recognition using convolutional neural architecture. The authors in [ 15 ], demonstrated an early application of recurrent neural networks for modeling scene text using orientation features.…”