Machine Learning and Image Processing for Character Recognition: An Application in Handwriting Instruction
International Research Journal of Modernization in Engineering Technology and Science · 2023
This research paper presents an innovative application of machine learning and image processing techniques for teaching handwriting skills.The proposed system utilizes a convolutional neural network (CNN) to accurately recognize characters written by students and provides valuable feedback on their handwriting.To enhance the input images' quality, an image processing algorithm incorporating edge detection, binarization, and noise reduction is employed.The system's potential applications extend to educational settings, where it can offer personalized feedback on students' handwriting skills, and to individuals seeking to improve their handwriting abilities through systematic practice.Feedback is provided on crucial aspects such as letter formation, size, spacing, and slant, enabling users to address specific areas for improvement.Experimental results demonstrate that the proposed system achieves high recognition accuracy for both uppercase and lowercase letters.Additionally, the system successfully identifies errors in input images, enabling targeted feedback to highlight areas requiring improvement.Leveraging recent technological advances, this research emphasizes the use of tablets as a tool to foster independent learning while respecting the perspectives, knowledge, and preferences of teachers.