“…A slight variation of this prediction scheme is also proposed in [33]. In [34], the context pixels are selected in such a way that the two boundary lines of the current pixel and context pixels form an obtuse angle. PPVO-based prediction is applied using such a context.…”