Embodiments described herein provide approaches for controlling inadvertent inputs to a mobile device. Specifically, at least one approach includes: detecting an operating mode of a mobile device by determining if a user is currently interacting with the mobile device; detecting an operating environment of the mobile device; receiving an input resulting from a physical gesture to an input area of a mobile device; comparing the input to a past history of inputs received by the mobile device; and determining whether the physical gesture is intended by the user based on the operating mode of the mobile device, the operating environment of the mobile device, and the past history of inputs (e.g., per device application). In one approach, an input controller selects logic to be applied in processing gestures based on a combination of user customization, interaction history, and environment characteristics. The selected logic is applied to subsequent gestures.