A method for handwritten word spotting based on particle swarm optimisation and multi‐layer perceptron

Abstract

This study presents a new method for handwritten keyword spotting. The innovation in this paper is to provide a model based on neural network architecture and an output based on the margin. At first, a neural network is designed such that its output determines whether a test word as an input is spotted or rejected. The intended neural network has one input layer, two middle layers, and one output layer. Another innovation in this study is optimising neural network weights based on swarm optimisation method. This optimisation model is used to train the neural network, so that the output has adequate margin for classification. The new components of the proposed classifier include new particle coding and new fitness function. Two layers are considered in coding particle, one for activating and deactivating neural network nodes and the other layer for acquiring proper values for weights. Different experiments with variety of parameters were designed for the multi-layer perceptron neural network. The experiments on three datasets: AMA Arabic dataset, IAM English dataset, and IFN/Farsi dataset yielded 83, 77, and 69% values, respectively, in the best condition. The results demonstrate that the proposed method has been better than the previous ones.

Hough Transform-Based Angular Features for Learning-Free Handwritten Keyword Spotting

“…In this work, the similarity score has been estimated using the confidence score from the Support Vector Machine (SVM) classifier. Recently, the use of local gradient and Gabor-based features is found in the work proposed by Tavoli and Keyvanpour [ 27 ] for KWS in document images written in Roman and Arabic scripts. They use particle swarm optimization that includes multilayer perceptron (MLP) as a learner.…”

“…Here, it is worth mentioning that distance and sequence matching-based techniques have been widely used in literature. Additionally, for performing KWS, in this work, we considered A different set-up of experiments could be found in the works [4,[26][27][28][29] where the use of multiple copies of image samples for each keyword could be found. Such set-ups are designed for a pre-defined set of keywords that need to be searched from a pool of words.…”

“…A different set-up of experiments could be found in the works [ 4 , 26 , 27 , 28 , 29 ] where the use of multiple copies of image samples for each keyword could be found. Such set-ups are designed for a pre-defined set of keywords that need to be searched from a pool of words.…”

“…In other words, the objective of these works is to decide whether a target word image belongs to the set of keywords (and if yes, then which one?). Multiple samples concerning each keyword used need to be prepared to create a trained module in a holistic way [ 4 , 26 , 27 ], or for tuning the parameters of the matching technique [ 28 , 29 ]. However, experimental set-ups mentioned in these works might not be possible in practical scenarios since such a system needs to be trained again if one wants to include a new keyword in the existing keyword set.…”

Vietnamese character recognition based on CNN model with reduced character classes

“…Text recognition is the process of converting the text in one or more images into documents such as those written on a computer [1][2][3]. Currently, the word recognition problem has been studied and widely used in the automation of office operations in many languages like English [4,5], Chinese [6,7], Japanese [8]. In Vietnam, there are some identification systems such as VietOCR software [9] and VNDOCR [10].…”