You have not specify what product you use. ABBYY technologies are enable to split the facing pages automatically but different products behave differently by default.
For example, ABBYY FineReader Engine 11 has the SplitType property of the PageSplittingParams object which specifies the type of the page split which should be performed. By default, the value of this property is PST_None, which means that a page should not be split. If the page you are going to process contains a dual-page spread use the PST_DoublePageSplit value for the SplitType property.
As for ABBYY Cloud OCR SDK, it cannot split a dual-page spread and this feature is not added to the service API, so there is no possibility to change this behavior.
This answer is marked "community wiki".
answered 06 May '15, 13:20
Oksana Serdyuk ♦♦