Page Splits Across "Low Density" Content

Pages may be split where there are fewer, non-blank pixels, broadly distributed across the horizontal line. For example, on music scores pages are often split across the middle of the grand staff where the non-blank content of the horizontal line may be just the vertical slice of the bracket on the left side of the page and a few, thin, vertical measure dividing lines through the rest of the scanned line. Similarly, tops or bottoms of the staves may include arcs, indicating phrases, ties, or slurs. These arcs may appear as only two, small marks on the scanned vertical line, yet it would be preferable to avoid splitting the page across them.

Can the sensitivity of the content gap identifier be adjusted, ideally as a command line argument, to provide for better identification of thinly distributed content?