OCR Handwriting Data Fine-Tuning and Labeling
ImageFine Tuning
Led the implementation of an end-to-end Optical Character Recognition (OCR) system using transformer models, focusing on handwriting data from the IAM Handwriting Dataset. Performed data preprocessing and model fine-tuning for improved handwritten text recognition. Oversaw the data labeling workflow as part of supervised fine-tuning of vision-encoder-decoder models.
• Handled image preprocessing tasks like resizing, padding, and normalization.
• Managed ground-truth labeling and text decoding for OCR evaluation.
• Set up evaluation with metrics like WER and CER for quality assurance.
• Enabled extensibility for real-world deployment of handwriting recognition models.
Led the implementation of an end-to-end Optical Character Recognition (OCR) system using transformer models, focusing on handwriting data from the IAM Handwriting Dataset. Performed data preprocessing and model fine-tuning for improved handwritten text recognition. Oversaw the data labeling workflow as part of supervised fine-tuning of vision-encoder-decoder models. • Handled image preprocessing tasks like resizing, padding, and normalization. • Managed ground-truth labeling and text decoding for OCR evaluation. • Set up evaluation with metrics like WER and CER for quality assurance. • Enabled extensibility for real-world deployment of handwriting recognition models.
2025 - 2025