Scholarlink

Segmentation of handwritten words into individual characters is an important step in developing handwritten optical character recognition. This paper provides an improved Bangla handwritten character segmentation algorithm using vertical filling operation. To make the system more accurate, the whole system is divided into two phases: pre-processing and segmentation. The system is implemented on a subset of the standard dataset ‘ICFHR2022 Competition on Indic Handwriting Text Recognition (IHTR 22)’. The system shows a promising result which indicates the efficiency of the suggested approach.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Simple Approach Towards Segmentation of Characters from Offline Handwritten Bangla Words

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Ankan Bhaduri,
  • Piyush Banka,
  • Abhishek Gupta,
  • Payel Rakshit,
  • Chayan Halder

摘要

Segmentation of handwritten words into individual characters is an important step in developing handwritten optical character recognition. This paper provides an improved Bangla handwritten character segmentation algorithm using vertical filling operation. To make the system more accurate, the whole system is divided into two phases: pre-processing and segmentation. The system is implemented on a subset of the standard dataset ‘ICFHR2022 Competition on Indic Handwriting Text Recognition (IHTR 22)’. The system shows a promising result which indicates the efficiency of the suggested approach.