Handwriting Recognition

50 papers with code • 3 benchmarks • 20 datasets


Use these libraries to find Handwriting Recognition models and implementations

Most implemented papers

LSTM: A Search Space Odyssey

flukeskywalker/highway-networks 13 Mar 2015

Several variants of the Long Short-Term Memory (LSTM) architecture for recurrent neural networks have been proposed since its inception in 1995.

OrigamiNet: Weakly-Supervised, Segmentation-Free, One-Step, Full Page Text Recognition by learning to unfold

IntuitionMachines/OrigamiNet CVPR 2020

On IAM we even surpass single line methods that use accurate localization information during training.

Speech Recognition with Deep Recurrent Neural Networks

HawkAaron/warp-transducer 22 Mar 2013

Recurrent neural networks (RNNs) are a powerful model for sequential data.

Full Page Handwriting Recognition via Image to Sequence Extraction

kingyiusuen/image-to-latex 11 Mar 2021

We present a Neural Network based Handwritten Text Recognition (HTR) model architecture that can be trained to recognize full pages of handwritten or printed text without image segmentation.

Multi-Dimensional Recurrent Neural Networks

philipperemy/tensorflow-multi-dimensional-lstm 14 May 2007

Recurrent neural networks (RNNs) have proved effective at one dimensional sequence learning tasks, such as speech and online handwriting recognition.

Spatially-sparse convolutional neural networks

facebookresearch/SparseConvNet 22 Sep 2014

Convolutional neural networks (CNNs) perform well on problems such as handwriting recognition and image classification.

A Critical Review of Recurrent Neural Networks for Sequence Learning

junwang23/deepdirtycodes 29 May 2015

Recurrent neural networks (RNNs) are connectionist models that capture the dynamics of sequences via cycles in the network of nodes.

ScrabbleGAN: Semi-Supervised Varying Length Handwritten Text Generation

amzn/convolutional-handwriting-gan CVPR 2020

This is especially true for handwritten text recognition (HTR), where each author has a unique style, unlike printed text, where the variation is smaller by design.

Segmental Recurrent Neural Networks

ykrmm/TREMBA 18 Nov 2015

Representations of the input segments (i. e., contiguous subsequences of the input) are computed by encoding their constituent tokens using bidirectional recurrent neural nets, and these "segment embeddings" are used to define compatibility scores with output labels.

Trainable Spectrally Initializable Matrix Transformations in Convolutional Neural Networks

NarayanSchuetz/SpectralLayersPyTorch 12 Nov 2019

In this work, we investigate the application of trainable and spectrally initializable matrix transformations on the feature maps produced by convolution operations.