1) The document discusses curvelet transformation and its application to object tracking. Curvelet transformation is a multiscale directional transform that can efficiently represent objects with curved edges using only a small number of coefficients. 2) It describes the stages of curvelet transformation including sub-band decomposition, smooth partitioning, renormalization, and ridgelet analysis. It also discusses the fast discrete curvelet transform implementation using unequally spaced fast Fourier transforms. 3) The proposed algorithm calculates the curvelet coefficients of frames to track objects based on the difference in curvelet energy between frames. Preliminary results on sample video frames are shown to demonstrate the calculation of curvelet coefficients.