1) The document outlines the steps in a digitization workflow for historical documents, including image enhancement, segmentation, and optical character recognition (OCR). 2) It describes hybrid methods for textline and word segmentation that do not require character recognition, making them suitable for documents with non-dictionary words. These include a connected component clustering approach and density-based projection profiling. 3) The methods are evaluated on historical documents, demonstrating accurate textline segmentation on over 2700 lines and word segmentation on over 14500 words.