Particle tracking, the analysis of individual moving elements in time series of microscopic images, enables burgeoning new applications, but there is need to better resolve conformation and dynamics. Here we describe the advantages of Delaunay triangulation to extend the capabilities of particle tracking in three areas: (1) discriminating irregularly shaped objects, which allows one to track items other than point features; (2) combining time and space to better connect missing frames in trajectories; and (3) identifying shape backbone. To demonstrate the method, specific examples are given, involving analyzing the time-dependent molecular conformations of actin filaments and lambda-DNA. The main limitation of this method, shared by all other clustering techniques, is the difficulty to separate objects when they are very close. This can be mitigated by inspecting locally to remove edges that are longer than their neighbors and also edges that link two objects, using methods described here, so that the combination of Delaunay triangulation with edge removal can be robustly applied to processing large data sets. As common software packages, both commercial and open source, can construct Delaunay triangulation on command, the methods described in this paper are both computationally efficient and easy to implement.