In this paper we have tried to focus on problems faced by visually impaired students of Orissa in school and public examinations and discussed the emerging solutions to those problems at the end in order to minimize those problems. Assessment plays an important role in the education system for evaluating student’s skill in that area. Assessment can be carried out by allowing students to appear exam in terms objective or non-objective type questions but due to the advantages offered by objective type assessment like offer speedier marking, reduced need for cross marking, greater fairness (reliability), more coverage of the syllabus and finally automate the marking process which enforces examination system to be held as objective rather non-objective type. Therefore, it is mostly implemented in examination system conducted in public or school exams. But it is very regret to mention that this assessment procedure is not well suitable for students those are visually impaired as they are unable to read what they have written or it is not possible to know if the ribbon has run out or if the line spacing is faulty unless the supervisor tells them. These are the main obstacles faced by visually challenged people which lead to disqualification for higher studies that ultimately reason to failure of motto i.e. “Education for All”. In this research study we have considered close ended objective type question. All possible responses of the closed questions are designed such a way that they are mutually exclusive. The closed ended questions are of five types namely fill-in-the-blanks questions, dichotomous questions, ranking scale questions, multiple choice questions and rating scale questions. As all the answers to the above types of questions carry isolated answers, we have developed Oriya isolated speech recognition system which can easily convert isolated answers uttered by visually impaired students to isolated text(isolated speech to isolated text). Hence, visually impaired students can answer the question as normal student because they do not need at all their hand and eyes to respond the answer. As a result of which overcomes the problem faced by visually challenged people to answer the questions without help of keyboard. Oriya isolated recognition system is developed using hidden Markov Model by incorporating1800 isolated Oriya words collected from 30 speakers in training stage. Testing stage is carried out by considering 5 speakers which yields 76.23% word accuracy on seen data and 58.86% on unseen data.
Mohanty, Sanghamitra and Swain, Basanta Kumar
"Markov Model Based Oriya Isolated Speech Recognizer-An Emerging Solution for Visually Impaired Students in School and Public Examination,"
International Journal of Computer and Communication Technology: Vol. 3
, Article 10.
Available at: https://www.interscience.in/ijcct/vol3/iss1/10