In papers the terms object pose estimation and object tracking are used side by side with different meanings. Can somebody explain me the difference?
"Pose" is a commonly used term that means position and orientation, where object tracking would just care about position but not necessarily the orientation.
So, for a vehicle on the street, you would want to know its pose so you could determine both where it is and if it is facing you or not.
For other applications, position alone is sufficient; think objects on a ballistic trajectory like basketballs, baseballs, etc. There the orientation doesn't matter.
When someone is referring to object tracking, he most likely talks about estimating the location of an object in an ordered sequence of images (e.g. a video). In this case, for most images, you already know where to expect an object (namely close to its location in the last image) and can use several detections to estimate the speed so that your prediction gets better. The location in tracking can be the region in the image in which the object is visible or even the 6d pose of the object. Main problems in tracking are occlusions or several similar objects in the scene. (think about a person walking through a group of other people)
Pose estimation is rather referring to computing the pose given a single image, in most cases with prior knowledge (e.g. texture or shape) of the object. Pose estimation can be used while tracking an object.