arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.07787cs.CVcs.LG
收藏

重新思考文本行识别模型

Rethinking Text Line Recognition Models

Google Research(谷歌研究院)

Daniel Hernandez Diaz, Siyang Qin, Reeve Ingle, Yasuhisa Fujii, Alessandro Bissacco

摘要:

本文研究了文本行识别问题。与大多数针对特定领域(如场景文本或手写文档)的方法不同,我们研究了开发一个通用架构的一般问题,该架构可以从任何图像中提取文本,而不管源或输入方式如何。我们考虑两个解码器系列(连接时间分类和Transformer)和三个编码器模块(双向LSTM,自我注意,和GRCL),并进行广泛的实验,以比较其准确性和性能广泛使用的场景和手写文本的公共数据集。我们发现,到目前为止,在文献中很少受到关注的组合,即自注意编码器与CTC解码器相结合,当与外部语言模型相结合并在公共和内部数据上进行训练时,在准确性和计算复杂性方面优于所有其他人。与更常见的基于Transformer的模型不同,该架构可以处理任意长度的输入,这是通用线路识别的要求。使用从多个来源收集的内部数据集,我们还暴露了当前公共数据集在评估线识别器的准确性方面的局限性,因为相对较窄的图像宽度和序列长度分布不允许观察Transformer方法在应用于长线转录时的质量下降。

英文摘要:

In this paper, we study the problem of text line recognition. Unlike most approaches targeting specific domains such as scene-text or handwritten documents, we investigate the general problem of developing a universal architecture that can extract text from any image, regardless of source or input modality. We consider two decoder families (Connectionist Temporal Classification and Transformer) and three encoder modules (Bidirectional LSTMs, Self-Attention, and GRCLs), and conduct extensive experiments to compare their accuracy and performance on widely used public datasets of scene and handwritten text. We find that a combination that so far has received little attention in the literature, namely a Self-Attention encoder coupled with the CTC decoder, when compounded with an external language model and trained on both public and internal data, outperforms all the others in accuracy and computational complexity. Unlike the more common Transformer-based models, this architecture can handle inputs of arbitrary length, a requirement for universal line recognition. Using an internal dataset collected from multiple sources, we also expose the limitations of current public datasets in evaluating the accuracy of line recognizers, as the relatively narrow image width and sequence length distributions do not allow to observe the quality degradation of the Transformer approach when applied to the transcription of long lines.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑