Using Wavelets for Compression and Multiresolution Search with Active Appearance Models

Active Appearance Models (AAMs) provide a method of modelling the appearance of objects in images and locating them automatically. Although the AAM approach is computationally eÆcient, the models used to search unseen images for the structures of interest are large typically the size of 100 images. This is perfectly practical for most 2-D images, but is currently impractical for 3-D images. We present a method for compressing the model information using a wavelet transform. The transform is applied to a set of training images in a shapenormalised frame, and coeÆcients of low variance across the training set are removed to reduce the information stored. An AAM is built from the training set using the wavelet coeÆcients rather than the raw intensities. We show that reliable image interpretation results can be obtained at a compression ratio of 20:1, which is suÆcient to make 3-D AAMs a practical proposition.