A speech-based driver assisting module for Intelligent Transport System

TL;DR

This paper presents a novel speech-based driver assistance system that converts roadside traffic panel images into speech to aid drivers, utilizing image detection, text extraction, and speech synthesis modules, supported by a new traffic panel corpus.

Contribution

It introduces a new approach in driver assistance by transforming traffic panel images into speech, including a custom corpus for Indian languages.

Findings

Successfully developed a three-module system for text detection, extraction, and speech conversion.

Created a corpus of 250 traffic panel images in Indian languages.

Demonstrated system effectiveness in assisting drivers with roadside traffic information.

Abstract

Aim of this research is to transform images of roadside traffic panels to speech to assist the vehicle driver, which is a new approach in the state-of-the-art of the advanced driver assistance systems. The designed system comprises of three modules, where the first module is used to capture and detect the text area in traffic panels, second module is responsible for converting the image of the detected text area to editable text and the last module is responsible for transforming the text to speech. Additionally, during experiments, we developed a corpus of 250 images of traffic panels for two Indian languages.