Denoising of natural images is the fundamental and challenging research problem of Image processing. This problem appears to be very simple however that is not so when considered under practical situations, where the type of noise, amount of noise and the type of images all are variable parameters, and the single algorithm or approach can never be sufficient to achieve satisfactory results. Fourier transform method is localized in frequency domain where the Wavelet transform method is localized in both frequency and spatial domain but both the above methods are not data adaptive. This thesis reviews the existing denoising algorithms, such as principal component analysis (PCA), Adaptive principal component analysis, and independent component analysis (ICA), and performs their comparative study with their parameters. Different types of noise can be removed by using these techniques, but in our project we use the Gaussian noise. Selection of the denoising algorithm is application dependent. Hence, it is necessary to have knowledge about the noise present in the image so as to select the appropriate denoising algorithm.