Monotonicity of optimized quantum $f$-divergence

We give an alternative proof for the data processing inequality of the optimized quantum $f$-divergences introduced in [Wil18] and extend the monotonicity to maps which are not completely positive.