Matjaz Jogan, PhD
We developed algorithms and representations for autonomous mapping and visual localization of mobile robots from omni-directional images. We also developed a purely emergent hierarchical mapping algorithm based on “recover and select” that creates maps of wide and unstructured environments from local appearance and odometry.
State of the art methods for visual categorization typically use learned feature based representations to discriminate between object categories. We propose an alternative approach based on hierarchical search for similarities to a prototype by synchronous binding of structural and appearance matches.
We introduce a method for an incremental updating of the popular subspace representation calculated by incremental Principal Component Analysis (IPCA). We also developed a method for efficient PCA of rotated templates without the decomposition of the covariance matrix.