Document Image Retrieval: Issues and Future Directions

Document Image Retrieval: Issues and Future Directions

The content discusses document image retrieval, which is a fascinating and well-liked field of study that has advanced through various stages. Recent developments indicate that it is still a growing area, largely due to the widespread use of document scanner apps on mobile and handheld devices. The paper aims to offer an understanding of document retrieval and classification, presenting a general architecture, highlighting associated challenges, and identifying new research prospects within the domain. Document Image Retrieval: Issues and Future Directions

Best OCR Android App

consume text material by the visually impaired and blind but braille forms of most books and educational material are not available especially in third world countries. For displaying digital text documents in braille form, electronic refreshable displays, have been available for quite some time. However since inception, they have been very expensive and out of reach financially for many in third world countries. This paper presents a solution to these problems – an open source text to braille scanner along with a unique low cost refreshable braille display. The proposed display generates all Braille characters by simple sliding arrangement of two eleven-slot pre-protruded sliders utilizing permutations of the slots. Unlike conventional piezo-electric based displays that consumes a large amount of power to generate each character by dynamically actuating various slots, the proposed display requires extremely low power just to carry out linear sliding of the sliders without the need of individual slot excitation. The text scanner is presented in the form of a mobile app, built using Google’s open source optical character recognition (OCR) engine. The scanner app is currently capable of converting English text printed in a multitude of fonts into braille in real time without any internet connectivity. The app also features assistive technologies such as voice guidance and tactile feedback to assist the visually impaired user in scanning documents. The proposed device is implemented in a modular fashion such that one can easily expand to have more characters to be displayed. Finally the performance of the proposed display device has been tested under various conditions and a very satisfactory performance is achieved, in terms of cost, comfort of use and speed.

We attempt to find ways for people who are paralyzed and are in wheelchairs to be able to use a scanner, but the same way everyone can use them. As some individuals may be paralyzed from the neck down, we build a system for them to still be able to scan documents and be part of the office workplace. We employed open-source computer vision techniques in order to develop and design the scanner system software. In addition, machine learning based text recognition open-source tool called Tesseract are used for the OCR module of the system. For the hardware implementation, we mount the components onto a headband including the microcontroller, camera, a flash circuit, and a power supply circuit to charge the battery. The camera component is developed to be able provide a live video feed instead of needing to create a database where the data will be stored. The live video is transmitted wirelessly trough microcontroller to a smart phone application. This application can work on both Apple IOS and Android operating system, and was built using Flutter, a crossplatform app development software using a unique language named Dart. The proposed system can improve the working conditions of the disabled people by providing them an assistive technology to participate in the office workplace.

Usually there are several solutions to any given problem. But all solutions are not equal. One solution might be compact and portable, but at the compromise of cost and vice versa. By limiting the resources available to only those that are low cost, we were forced to come up with a simple cost effective solution to the problem at hand. With an expected cost of the total product being less than 100, we are very hopeful that people who can really benefit from such assistive technologies
will be able to afford them as well.