Wavelets decompose signals into different frequency bands using scaling functions and wavelet basis functions. The discrete wavelet transform (DWT) uses these functions to represent a discrete signal as a sum of wavelet coefficients at different scales. The fast wavelet transform provides an efficient algorithm to compute the DWT by successively applying filters and downsampling. Wavelet packets generalize the DWT by allowing decomposition of both low and high frequency bands at each level. Lifting transforms provide an alternative method to construct wavelets by splitting, predicting, and updating the signal. Two-dimensional wavelets extend the concepts to images by applying separable 1D wavelets along rows and columns.