Figure 2

Detection of ( L , d , r )-repeats and two overlapping parallelograms. Two parallelograms that overlap. The dark grey parallelogram in the figure detects q-hits between w = s[a, a + L - 1] and any word w _k = s[i, j] with i ∈ [c + a, c + d + a] and j ∈ [c + a + L - 1, c + d + a + L - 1]. The word z = s[c + a, c + d + a + L - 1] of length L + d contains the word w _k which in turn contains the word x = s[c + d + a, c + a + L - 1] of length L - d. Analogously for the light gray parallelogram, the word z' = s[c' + a, c' + d + a + L - 1] of length L + d contains a word ${w^{\prime }}_k$ which in turn contains the word x' = s[c' + d + a, c' + a + L - 1] of length L - d. The words w _k and ${w^{\prime }}_k$ are not shown because their length is variable. They necessarily overlap because they both contain the word p = s[c' + d + a, c + a + L - 1], that is the overlap between x and x'.