Handwriting Text

Handwriting text recognition (HTR)

• Baoguang Shi, Xiang Bai, and Cong Yao. An endto- end trainable neural network for image-based sequence recognition and its application to scene text recognition. IEEE transactions on pattern analysis and machine intelligence, 39(11):2298–2304, 2016. -> using convolutional recurrent neural network (CRNN) architecture for scene text recognition

• Arik Poznanski and Lior Wolf. Cnn-n-gram for handwriting word recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2305–2314, 2016. -> using an RCNN to estimate an n-gram profile of an image and match it to the profile of an existing word from a dictionary

• Sebastian Sudholt and Gernot A Fink. Phocnet: A deep convolutional neural network for word spotting in handwritten documents. In 2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), pages 277–282. IEEE, 2016. -> improves the architecture from Poznanski et al 2016

• Jorge Sueiras, Victoria Ruiz, Angel Sanchez, and Jose F Velez. Offline continuous handwriting recognition using sequence to sequence neural networks. Neurocomputing, 289:119–128, 2018. -> use attention decoder instead of RCNN outputs

• Kartik Dutta, Praveen Krishnan, Minesh Mathew, and CV Jawahar. Improving cnn-rnn hybrid networks for handwriting recognition. In 2018 16th International Conference on Frontiers in Handwriting Recognition (ICFHR), pages 80–85. IEEE, 2018. -> they mix several strategies developed in previous years into one architecture

• Jeonghun Baek, Geewook Kim, Junyeop Lee, Sungrae Park, Dongyoon Han, Sangdoo Yun, Seong Joon Oh, and Hwalsuk Lee. What is wrong with scene text recognition model comparisons? dataset and model analysis, 2019. -> advances in text scene recognition

Handwriting text generation (HTG)

• Alex Graves. Generating sequences with recurrent neural networks. arXiv preprint arXiv:1308.0850 2013. -> LSTM for generating text sequences

• Bo Ji and Tianyi Chen. Generative adversarial network for handwritten text. arXiv:1907.11845, 2019 -> extends Graves 2013 with a Discriminator and converts it into a GAN paradigm

• Emre Aksan, Fabrizio Pece, and Otmar Hilliges. Deepwriting: Making digital ink editable via deep generative modeling. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems, pages 1–14 2018. -> aka DeepWriting; introduces control of the styling for Graves 2013.

• Eloi Alonso, Bastien Moysset, and Ronaldo Messina. Adversarial generation of handwritten text images conditioned on sequences. arXiv preprint arXiv:1903.00277, 2019 -> generate text images conditioned on input text; for this, they extend a BigGAN architecture (Brock et al. 2018); the generator can produce images of fied size but containing text of variable length

• Sharon Fogel, Hadar Averbuch-Elor, Sarel Cohen, Shai Mazor, Roee Litman. ScrabbleGAN: Semi-Supervised Varying Length Handwritten Text Generation. arXiv:2003.10557, 2020 -> build upon Alonso et al 2019 to produce a network that generates images with both variable text length and variable image size.