Two New Techniques for Evaluating Trackers in SynthEyes
In this tutorial, Russ Andersson demonstrates new techniques in SynthEyes 2011 for identifying problematic trackers and frames within a shot. The example shot features moving water and wind-blown leaves, which present tracking challenges. Starting with an auto track, Russ introduces two methods to identify problematic areas: the tracker radar display and the RMS error coloring script. The tracker radar display shows circles around trackers that indicate error levels, with larger circles representing higher errors. This visual aid helps quickly identify problematic areas in the shot, such as moving branches and grass. The RMS error coloring script assigns colors to trackers based on their overall error, with green for low error, yellow for moderate error, and red for high error. This method provides a clear visual distinction of stable and unstable trackers. Additionally, the script allows users to toggle between the error-based colors and the original tracker colors, preserving any custom color assignments. These tools help users efficiently identify and address problematic trackers, improving the overall tracking quality.