Systems and methods which operate to identify interventional instruments and/or other objects in images are shown. Embodiments operate to extract relevant information regarding interventional instruments from a multi-dimensional volume for presenting the information to a user in near real-time with little or no user interaction. Objects may be identified by segmenting a multi-dimensional volume, identifying a putative object of interest in multiple multi-dimensional volume segments, and determining a position of the object of interest within the multi-dimensional volume using the putative object of interest segment identifications. Identification of objects of interest according to embodiments may be utilized to determine an image plane for use in displaying the objects within a generated image, to track the objects within the multi-dimensional volume, etc., such as for medical examination, interventional procedures, diagnosis, treatment, and/or the like.
