In previous work we use B-spline wavelets to detect Gaussian objects in Markov noise. The authors now generalize the method and use it to adaptively enhance images according to the detector output. They examine a class of multiscale matched filters implemented using the wavelet transform, and show how these filters may be employed to detect objects in complex backgrounds. In two contrasting applications they use the approach to detect microcalcifications in mammograms and ships in infrared images.