This paper proposes Parameterized Kernel Principal Component Analysis (PKPCA), an extension of Parameterized Appearance Models (PAMs) that uses Kernel PCA (KPCA) for learning a non-linear appearance model invariant to rigid and/or non-rigid deformations.
Parameterized Appearance Models (PAMs) (e.g., eigen-tracking, active appearance models, morphable models) use Principal Component Analysis (PCA) to model the shape and appearance of objects in images. Given a new image with an unknown appearance/shape configuration, PAMs can detect and track the object by optimizing the model’s parameters that best match the image. While PAMs have numerous advantages for image registration relative to alternative approaches, they suffer from two major limitations: First, PCA cannot model non-linear structure in the data. Second, learning PAMs requires precise manually labeled training data. KPCA, an extension of PAMs that uses Kernel PCA (KPCA) for learning a non-linear appearance model invariant to rigid and/or non-rigid deformations. The authors demonstrate improved performance in supervised and unsupervised image registration and present a novel application to improve the quality of manual landmarks in faces. In addition, they suggest a clean and effective matrix formulation for PKPCA. (Published abstract provided)
Similar Publications
- Fingerprint Reconstruction: From Minutiae to Phase
- Temporal Arcs of Mental Health: Patterns Behind Changes in Depression over Time
- How the Work Being Done on Statistical Fingerprint Models Provides the Basis for a Much Broader and Greater Impact Affecting Many Areas within the Criminal Justice System