Just because I've never seen it described here, I'll mention one method I use to read image-based multi-column PDFs on my ancient IPAQ. It avoids having to do OCR and handles the multiple column format well.
I read it with uBook, in the comic reader mode. The comic reader mode is designed to automatically process an image by slicing it vertically or horizontally into strips, then slice those strips into smaller rectangular image chunks.
For a normal comic page, it autoslices the image horizontally first, then vertically cuts the horizontal comic strips into individual panes. Each individual pane is displayed sequentially and the small image chunks fill the screen, allowing text to be read easily.
For an image based multicolumn PDF, I let it autoslice vertically first to produce text columns. Then it slices each column horizontally between the paragraphs. The space between columns and the space between paragraphs is treated just like the space that frames the comic panes. The only difference is the slicing order.
|