Doctoral research · Generative AI & computer vision
Synthetic image evaluation for OCR
Evaluate generated images and use synthetic data in handwritten-digit recognition.
Generative modeling · Synthetic data · OCR
Doctoral research at Southern Illinois University Carbondale · Dissertation, 2023
The problem & approach
Investigate synthetic-image quality and data augmentation for Arabic handwritten-digit optical character recognition (OCR), connecting generative-model evaluation with recognition experiments.
Methods & contributions
- Compare conditional variational autoencoders and conditional generative adversarial networks.
- Develop the Synthetic Image Evaluation Procedure for assessing generated images.
- Study feature extraction and synthetic-image augmentation in OCR experiments.
Research connections
2024 · Preprint · Project publication
Advancing Generative Model Evaluation: A Novel Algorithm for Realistic Image Synthesis and Comparison in OCR System →
Publication record ↗ (opens in a new tab)2023 · Conference paper · Project publication
Generative Data Augmentation for Arabic Handwritten Digit Recognition Boosting Real-time OCR Capabilities →
Publication record ↗ (opens in a new tab)2023 · Dissertation · Project publication
A New Approach to Synthetic Image Evaluation →
Publication record ↗ (opens in a new tab)