Title | ||
---|---|---|
A Kernel-based l2 norm regularized least square algorithm for vehicle logo recognition |
Abstract | ||
---|---|---|
We consider the problem of automatically recognizing the vehicle logos from the frontal views with varying illumination, as well as certain corruption. To better address the problem, a kernel-based l2 norm regularized least square (RLS) algorithm is proposed in the paper. Kernel technique is smoothly combined with the l2 norm RLS algorithm to enhance the performance of vehicle logo recognition (VLR). As an extension, the improvement of dictionary is also considered. A simple mechanism of constructing an adaptive online dictionary has been presented and experimented. Experimental results show that our proposed algorithm outperforms the original l2 norm RLS algorithm and the l1 norm based algorithms. © 2014 IEEE. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1109/ICDSP.2014.6900742 | International Conference on Digital Signal Processing, DSP |
DocType | Volume | Citations |
Conference | 2014-January | 1 |
PageRank | References | Authors |
0.35 | 0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Weiyang Liu | 1 | 31 | 6.29 |
Yandong Wen | 2 | 115 | 8.27 |
Pan Kai | 3 | 4 | 4.47 |
Li Hui | 4 | 173 | 34.14 |
Zou Yuexian | 5 | 215 | 39.62 |