Point tracking
A seminal paper in Motion Tracking is that of Lucas and Kanade[1]. The original paper maybe hard to read for a beginner. You could perhaps go through a tutorial. Shi and Tomasi[2] advocated a improved metric for the features used for tracking and an affine motion model, as opposed to the translational model in the LK tracker.
Baker and Mathews[3] suggested a computational efficient tweak to the KLT tracker and established that it was the best-performing point tracking algorithm, 20 years after its conception.

Template based tracking
Then came a flurry of Machine Learning (ML) approaches that learnt parts of the object template or the entire template itself using SVM, AdaBoost etc to find the object in the test sequences. The recent paper by Boris Babenko[4], the survey of Online Object Tracking approaches[7], PROST[5], VoT (Visual Object Tracking Challenge)[6], are highly cited papers, not too far in the past.

Tracking by detection
Kalal et al[8] first proposed to pose the Object tracking problem as an object detection in every frame of the sequence, making the actual tracking trivial. Look up papers that cite this one for a whole range of techniques in this genre.

Deep Learning
Very recent approaches use Convolutional Neural Networks that in the learning, which gives brilliant performance. A recent paper by Jin et al[9] and the presentation on Deep Learning in OR, Segmentation and Detection will give you a generic idea of these methods.

References
[1] Lucas and Kanade 1981, An iterative image registration technique with an application to stereo vision
[2] Shi and Tomasi 1994, Good Features to track
[3] Baker and Mathews 2004, Lucas-Kanade 20 years in: A unifying framework
[4] Babenko, Yang and Belongie 2009, Visual Tracking with Multiple Instance Learning
[5] Santner et al 2010, PROST: Parallel robust online simple tracking
[6] The Visual Object Tracking 2014 challenge results: VoT 2014 and VoT 2013
[7] Wu, Lim and Yang 2013, Online Object Tracking
[8] Kalal, Mikolajczyk and Matas 2012, Tacking-Learning-Detection
[9] Jin et al, 2013, Tracking with deep Neural Networks