Paired Image to Image Translation for Strikethrough Removal From Handwritten Words

Transcribing struck-through, handwritten words, for example for the purpose of genetic criticism, can pose a challenge to both humans and machines, due to the obstructive properties of the superimposed strokes. This paper investigates the use of paired image to image translation approaches to remove strikethrough strokes from handwritten words. Four different neural network architectures are examined, ranging from a few simple convolutional layers to deeper ones, employing Dense blocks. Experimental results, obtained from one synthetic and one genuine paired strikethrough dataset, confirm that the proposed paired models outperform the CycleGAN-based state of the art, while using less than a sixth of the trainable parameters.

[1]  TexRGAN: a deep adversarial framework for text restoration from deformed handwritten documents , 2021 .

[2]  Lambert Schomaker,et al.  Automatic removal of crossed-out handwritten text and the effect on writer verification and identification , 2008, Electronic Imaging.

[3]  H. K. Chethan,et al.  A Study on Identification and Cleaning of Struck-Out Words in Handwritten Documents , 2021 .

[4]  D. Hulle.  The Stuff of Fiction: Digital Editing, Multiple Drafts and the Extended Mind , 2014 .

[5]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  Bidyut Baran Chaudhuri,et al.  Impact of struck-out text on writer identification , 2017, 2017 International Joint Conference on Neural Networks (IJCNN).

[7]  Jayanta Mukhopadhyay,et al.  Detection and Localisation of Struck-Out-Strokes in Handwritten Manuscripts , 2021, ICDAR Workshops.

[8]  Anders Hast,et al.  Strikethrough Removal from Handwritten Words Using CycleGANs , 2021, ICDAR.

[9]  Bidyut Baran Chaudhuri,et al.  An approach for detecting and cleaning of struck-out handwritten text , 2017, Pattern Recognit..

[10]  Jorge Calvo-Zaragoza,et al.  A selectional auto-encoder approach for document image binarization , 2017, Pattern Recognit..

[11]  Hala Neji,et al.  Adversarial Autoencoders for Denoising Digitized Historical Documents: The Use Case of Incunabula , 2019, 2019 International Conference on Document Analysis and Recognition Workshops (ICDARW).

[12]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[13]  James A. Thom,et al.  A deep learning approach to handwritten text recognition in the presence of struck-out text , 2019, 2019 International Conference on Image and Vision Computing New Zealand (IVCNZ).

[14]  Yoshua Bengio,et al.  The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[15]  Horst Bunke,et al.  The IAM-database: an English sentence database for offline handwriting recognition , 2002, International Journal on Document Analysis and Recognition.

[16]  Jean-Christophe Burie,et al.  Binarization Strategy Using Multiple Convolutional Autoencoder Network for Old Sundanese Manuscript Images , 2021, ICDAR Workshops.

[17]  N. Otsu.  A threshold selection method from gray level histograms , 1979 .

[18]  Palaiahnakote Shivakumara,et al.  A Connected Component-Based Deep Learning Model for Multi-type Struck-Out Component Classification , 2021, ICDAR Workshops.

[19]  Laurence Likforman-Sulem,et al.  HMM-based Offline Recognition of Handwritten Words Crossed Out with Different Kinds of Strokes , 2008 .

[20]  Ji-Rong Wen,et al.  Skip-Connected Deep Convolutional Autoencoder for Restoration of Document Images , 2018, 2018 24th International Conference on Pattern Recognition (ICPR).

[21]  Alexei A. Efros,et al.  Image-to-Image Translation with Conditional Adversarial Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[22]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.