Abstract:Handwriting recognition is one of the desirable attributes of document comprehension and analysis. It is concerned with the documents writing style and characteristics that distinguish the authors. The diversity of text images, notably in images with varying handwriting, makes the process of learning good features difficult in cases where little data is available. In this paper, we propose a novel scheme to identify the author of a document based on the input word image. Our method is text independent and does not impose any constraint on the size of the input image under examination. To begin with, we detect crucial components in handwriting and extract regions surrounding them using Scale Invariant Feature Transform (SIFT). These patches are designed to capture individual writing features (including allographs, characters, or combinations of characters) that are likely to be unique for an individual writer. These features are then passed through a deep Convolutional Neural Network (CNN) in which the weights are learned by applying the concept of Similarity learning using Siamese network. Siamese network enhances the discrimination power of CNN by mapping similarity between different pairs of input image. Features learned at different scales of the extracted SIFT key-points are encoded using Sparse PCA, each components of the Sparse PCA is assigned a saliency score signifying its level of significance in discriminating different writers effectively. Finally, the weighted Sparse PCA corresponding to each SIFT key-points is combined to arrive at a final classification score for each writer. The proposed algorithm was evaluated on two publicly available databases (namely IAM and CVL) and is able to achieve promising result, when compared with other deep learning based algorithm.

Bag of Features Approach for Offline Text-Independent Chinese Writer Identification

Off- Line Chinese Writer Identification Based on Character-Level Decision Combination

Chinese Handwriting Identification Based on Stable Spectral Feature of Texture Images

Offline Text-Independent Writer Identification Based on Scale Invariant Feature Transform

Offline text-independent writer identification using stroke fragment and contour based features

An Off-line Chinese Writer Retrieval System Based on Text-sensitive Writer Identification

Text-Independent Writer Identification Via CNN Features and Joint Bayesian

A Wavelet-Based Statistical Method for Chinese Writer Identification

A Contourlet-Based Method for Writer Identification

Character-level Chinese Writer Identification using Path Signature Feature, DropStroke and Deep CNN

Text-Independent Writer Identification Using Improved Structural Features.

Dual-Branch Multitask Fusion Network for Offline Chinese Writer Identification.

Chinese Handwriting-Based Writer Identiefication By Texture Analysis

A Novel Method for Offline Handwriting-Based Writer Identification

Writer Adaptation Using Bottleneck Features and Discriminative Linear Regression for Online Handwritten Chinese Character Recognition

Handwriting-based Personal Identification.

Writer Adaptation Via Deeply Learned Features for Online Chinese Handwriting Recognition

Writer Adaptive Feature Extraction Based on Convolutional Neural Networks for Online Handwritten Chinese Character Recognition

Robust End-to-End Offline Chinese Handwriting Text Page Spotter with Text Kernel

Siamese based Neural Network for Offline Writer Identification on word level data

Exploiting Multi-Scale Fusion, Spatial Attention and Patch Interaction Techniques for Text-Independent Writer Identification