We propose a new framework that extends the standard Probability Hypothesis\nDensity (PHD) filter for multiple targets having $N\\geq2$ different types based\non Random Finite Set theory, taking into account not only background clutter,\nbut also confusions among detections of different target types, which are in\ngeneral different in character from background clutter. Under Gaussianity and\nlinearity assumptions, our framework extends the existing Gaussian mixture (GM)\nimplementation of the standard PHD filter to create a N-type GM-PHD filter. The\nmethodology is applied to real video sequences by integrating object detectors'\ninformation into this filter for two scenarios. For both cases, Munkres's\nvariant of the Hungarian assignment algorithm is used to associate tracked\ntarget identities between frames. This approach is evaluated and compared to\nboth raw detection and independent GM-PHD filters using the Optimal Sub-pattern\nAssignment metric and discrimination rate. This shows the improved performance\nof our strategy on real video sequences.\n