Dyslexia handwriting detection using Convolutional Neural Network (CNN) algorithm / Sofea Najihah Mohd Zaki
Dyslexia is a neurological learning situation that affects a person's ability to read, spell, write, and speak. Children with dyslexia frequently suffer from negative emotions including sadness, anger, and low confidence. Thus, it's critical to detect dyslexic children early to provide the...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/96469/1/96469.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Dyslexia is a neurological learning situation that affects a person's ability to read, spell, write, and speak. Children with dyslexia frequently suffer from negative emotions including sadness, anger, and low confidence. Thus, it's critical to detect dyslexic children early to provide them proper support that they need. The handwriting style is a simple way to detect the most common dyslexia sign. Many computational and intelligent techniques have been put forward, and they provided a variety of outcomes to assess the recommended system's effectiveness. This project attempts to accurately detect the type of dyslexic handwriting. Convolutional Neural Network (CNN) algorithm was chosen as one of the possible solutions after a thorough analysis of many algorithms for dyslexic handwriting identification. This is because CNN can automatically extract new image data from written material. The CNN model is trained and evaluated on a wide range of handwritten sample datasets. It successfully detects dyslexia-related handwriting type (normal or reversal) with an 90% accuracy rate and high precision 0.93 for Class 1 (Reversal). The dyslexia handwriting detection prototype demonstrates how machine learning may improve the detect handwriting issues in educational settings. A user-friendly desktop prototype was developed, and users may upload handwritten samples and get immediate results for the dyslexia handwriting type (normal or reversal) and status of handwriting (detect or not detect). Further enhancements might involve including machine learning algorithms to improve the prototype's accuracy by learning from a larger dataset, which would eventually improve the prototype's ability to offer deep understanding into handwriting patterns related to dyslexia. |
|---|