Handwritten Mathematical Expression Recognition via Attention Aggregation based Bi-directional Mutual Learning
This work addresses the problem of improving recognition accuracy for handwritten mathematical expressions, which is important for educational and document digitization applications, and it is incremental by building on existing attention-based encoder-decoder models.
The paper tackles handwritten mathematical expression recognition by proposing an attention aggregation based bi-directional mutual learning network (ABM) that uses two inverse decoders with mutual distillation and multi-scale attention integration, achieving recognition accuracies of 56.85% on CROHME 2014, 52.92% on CROHME 2016, and 53.96% on CROHME 2019 without data augmentation or model ensembling.
Handwritten mathematical expression recognition aims to automatically generate LaTeX sequences from given images. Currently, attention-based encoder-decoder models are widely used in this task. They typically generate target sequences in a left-to-right (L2R) manner, leaving the right-to-left (R2L) contexts unexploited. In this paper, we propose an Attention aggregation based Bi-directional Mutual learning Network (ABM) which consists of one shared encoder and two parallel inverse decoders (L2R and R2L). The two decoders are enhanced via mutual distillation, which involves one-to-one knowledge transfer at each training step, making full use of the complementary information from two inverse directions. Moreover, in order to deal with mathematical symbols in diverse scales, an Attention Aggregation Module (AAM) is proposed to effectively integrate multi-scale coverage attentions. Notably, in the inference phase, given that the model already learns knowledge from two inverse directions, we only use the L2R branch for inference, keeping the original parameter size and inference speed. Extensive experiments demonstrate that our proposed approach achieves the recognition accuracy of 56.85 % on CROHME 2014, 52.92 % on CROHME 2016, and 53.96 % on CROHME 2019 without data augmentation and model ensembling, substantially outperforming the state-of-the-art methods. The source code is available in https://github.com/XH-B/ABM.