The problem of separating linear features from a textured background is of importance in many applications. It has been shown that the Fourier transform can be used in conjunction with polar transformation to "lift" linear features from the background texture. However, while the Fourier transform works well with lines that are spread throughout the entire image, it is less effective when the linear features are of varied length and thickness. We propose approaches based on a windowed Fourier transform and wavelet packet decomposition to lift randomly located lines of varied lengths and thickness. The reasoning underlying the development of the approaches is presented along with comparative examples.