فهرست نشریات علمی دسترسی آزاد ایران

Optical Character Recognition (OCR) is a technology that enhances the accessibility and inclusivity of learning materials for students across various age groups and skills. OCR can transform physical materials such as textbooks, research papers, and handwritten notes into digital text files. OCR systems optimized for structured text (e.g., printed books) may not perform well in real-world educational settings where handwritten notes, mixed scripts, and annotations are common. A direct performance comparison between OCRCHA and Deep Learning (DL) models on large, diverse datasets could highlight its advantages and limitations. This enhances the availability of study materials for students. The presence of varied handwriting styles and inadequate scan quality presents challenges for the OCR model. A subfield of artificial intelligence known as fuzzy logic has the potential to address these challenges, especially within the education sector. This work introduces Chaotic Fuzzy Logic (OCRCHA), a distinctive OCR system architecture designed for learning environments, integrating chaotic Large Language Models (LLM) with fuzzy logic. The system utilizes a feed-forward network to improve performance, integrating fuzzy logic with chaotic theory to increase flexibility and response accuracy. The experimental outcomes demonstrate notable performance improvements, highlighted by enhancements in F1 score, accuracy, and recall. This study introduces an enhanced iteration of the OCRCHA system to achieve more reliable character identification and demonstrates its efficacy in optical English letter recognition.