We present a learning method for classification using multiple manifold-valued features. Manifold techniques are becoming increasingly popular in computer vision since Riemannian geometry often comes up as a natural model for many descriptors encountered in different branches of computer vision. We propose a feature combination and selection method that optimally combines descriptors lying on different manifolds while respecting the Riemannian geometry of each underlying manifold. We use our method to improve object recognition by combining HOG  and Region Covariance  descriptors that reside on two different manifolds. To this end, we propose a kernel on the n-dimensional unit sphere and prove its positive definiteness. Our experimental evaluation shows that combining these two powerful descriptors using our method results in significant improvements in recognition accuracy.
|Title of host publication||2013 International Conference on Digital Image Computing|
|Subtitle of host publication||Techniques and Applications, DICTA 2013|
|Publication status||Published - 1 Dec 2013|
|Event||Digital Image Computing Techniques and Applications 2013 - Wrest Point Hotel, Hobart, Australia|
Duration: 26 Nov 2013 → 28 Nov 2013
Conference number: 15th
|Name||2013 International Conference on Digital Image Computing: Techniques and Applications, DICTA 2013|
|Conference||Digital Image Computing Techniques and Applications 2013|
|Abbreviated title||DICTA 2013|
|Period||26/11/13 → 28/11/13|
|Other||The International Conference on Digital Image Computing: Techniques and Applications (DICTA) is the main Australian Conference on computer vision, image processing, pattern recognition, and related areas. DICTA was established as a biennial conference in 1991 and became an annual event in 2007. It is the premier conference of the Australian Pattern Recognition Society (APRS).|
The conference will be held at Wrest Point Hotel, Hobart, Tasmania, Australia from 26 - 28 November 2013.