Reading is an essential skill for literacy development in children. But it is a challenge for children with dyslexia because of phonological-core deficits. Poor reading skills have an impact on vocabulary development and to exposure to relevant background knowledge. It affects the ability to interpret what one sees and hears or the ability to link information from different parts of the brain. Dyslexic children face many challenges in their educational life due to reading difficulty. Support to dyslexic children include computer-based applications and multi-sensory methods like text-to-speech and character animation techniques. Some applications provide immediate reading intervention facility. Automatic speech recognition (ASR) is a new platform with immediate intervention for assisting dyslexic children to improve their reading ability. Findings contribute to develop a suitable approach to correct the reading mistakes of dyslexic children. Speech recognition technology provides the most interactive environment between human and machine.
TopIntroduction
Dyslexia is a language based learning disability that affects reading. Dyslexia is most common reading difficulty, it is estimated that one out of 10 children is dyslexic. According International Dyslexia Association 15 to 20% world population detected symptom of dyslexia. Dyslexia has created a problem in processing word-sounds, identify the sound of letters. They have problem in understanding reversal letters and confusion between letters. Result of dyslexia phonics problems in language; reduce reading capability, vocabulary growth, and other knowledge. Table 1 represents some assistive software for dyslexic children to enhance learning ability. Dyslexia may impede child’s developing and learning capability. The performance of a dyslexic child is very low: trouble in reading at a good pace without mistake.
Due to difficulty in dyslexic children mastering in reading is a challenge. Very limited research for dyslexic children has been till now in regional languages. We find that dyslexia still not widely known in India and the research on this topic is extremely scarce.
To facilitate learning experience for dyslexics children help is essential needed (Husni & Jamaludin, 2009). People with dyslexia fail to comprehend the text as presented. A dyslexic person does not understand what is written, gives miss-meaning of the passage, and forget the sequence of events (Rahman, et al. 2012). Dyslexia can be managed or sustained with the help of special methods or techniques. Developing computer technologies helps to parents and teachers of dyslexic children in their teaching with easy and efficient manner. Dyslexia mostly affects on school age population around 6% to 17%. FunLexia is an educational game to help dyslexia children to learn Arabic. A heuristic evaluation and speech therapists indicates that FunLexia is supportive tool and promote leaning for children with dyslexia (Ouherrou et al. 2018). Graphogame is a game help to children to read in their local language.
Speech recognition is providing confidence and timing a child to track during read- aloud. The speech recognizer must follow behaviors of students quickly and accurately, when students reading fluently the cursor appear at right time and at right place and pauses when students hesitate. The students receive feedback about their pronunciation (Hagen et al., 2004). Multimedia elements (Sampathkumar et al 2020) are used as guidelines for learning skills in dyslexic students through special instructions.
Assist to dyslexic learners in learning within normal peer group needs an adaptive reading assistance system that is easy to understand. Adaptive reading assistance (Reyana, & Kautish, 2021) allows dyslexic students to reading text documents arbitrary using augmented way such as emphasis on characters, text highlighting, segmenting words. AGENT-DYSL system is adaptive to presentation of individual student and suggest individual for further training (Schmidt & Schneider, 2007).
Watch Me! a reading software designed for young children provide reading practice, sense of reading and comprehension awareness. Speech recognition used to assess students reading ability and gives individual feedback (Williams et al., 2013). ASR providing a reading tutor towards advancing that listen what is read by dyslexic children and give correct or incorrect reading feedback. Dyslexic children read aloud 114 words in Bahasa Melayu and speech recorded. ASR trained based on resulting utterances that transcribed and the pronunciations are modeled (Husni & Jamaluddin, 2008).
Recurrent neural network (Rani, & Kautish, 2018) used for creating speech recognition in Bengali language to find character probabilities. There has been widespread use of speech recognition at beginning of twenty-first century. It is very helpful for persons who have typing difficulty and those with spelling difficulties. Two models for Bengali speech recognition system, first model for recognize spoken language using Convolutional Neural Network (CNN) and second model use recurrent neural network. HMM phoneme and word based models used for training isolated and continuous word speech recognition. The CNN model accuracy convert speech to text is 86.058% and 30000 Bengali words used to train RNN model which converts speech signal into text transcriptions (Islam et al., 2019).