Real-Time Scale Invariatnt Object and Face Tracking Using Gabor Wavelet Templates

This paper describes a real-time technique for scale invariant object or face tracking with standard PC hardware. The tracking method is based on a low redundancy (compressed) object image represcntation. For image decomposition a fast non-iterative transform based on the odd-symmetric gabor functions is used, which guarantees on the one hand a low redundancy of the resulting representation and on the other hand an automatic detection of the significant features in the image. Tracking control is realized by a scale factor discriminator control technique and convolution based 2D cross-correlation. Experiments show that this method enables real-time object or face tracking in a wide scale changing range and provides strong robustness against camera shaking, which is important for visual robot contro.