Modified butterfly structure for efficient implementation of pruned fast cosine transform

In the computation of pruned fast cosine transforms (FCTs), overheads are unavoidable in order to keep track of the complete and incomplete butterflies in each stage. The authors present a modification to the FCT butterfly structure proposed to reduce these overheads. Using this modified structure, the computation of the coefficients can be stopped at any point and resumed from the same point without any overhead. This structure also performs well for complete computation of FCTs