Multi-band modelling of appearance

Abstract

Earlier work has demonstrated generative models capable of synthesising near photo-realistic grey-scale images of objects. These models have been augmented with colour information, and recently with edge information. This paper extends the Active Appearance Model framework by modelling the appearance of both derived feature bands and an intensity band. As a special case of featureband augmented appearance modelling we propose a dedicated representation with applications to face segmentation. The representation addresses a major problem within face recognition by lowering the sensitivity to lighting conditions. Results show that localisation accuracy of facial features is considerably increased using this appearance representation under diffuse and directional lighting and at multiple scales.

Feature-Based Lucas–Kanade and Active Appearance Models

“…Similar features are employed in [21] to create a robust similarity optimization criterion. In [27], the intensities appearance model is replaced by a mixture of grayscale intensities, hue channel and edge magnitude.…”

Appearance capturing and animation techniques for realistic, true-identity gaming avatars

“…We use homemade ZigBee-USB dongle, which embed a XBee module from Digi [10]. Whereas the wireless communication is not necessary in the standalone version, it still allows the system to be isolated from the main power line and thus, to be electrically safe for the user (the sensors are battery powered).…”

Filtered Component Analysis to Increase Robustness to Local Minima in Appearance Models

“…al [5] found that a non-linear representation of edge structure could improve subspace matching. Stegmann and Larsen [24] report that building subspaces for AM in an augmented space of intensity, hue and edges performed better in the task of localizing faces. In similar fashion, [7] make use of wedgelet regression trees to reduce the computational complexity of standard Active AM.…”