© The Institution of Engineering and Technology
Due to the heavy computation cost, the bilateral filter cannot be applied to real-time applications. To deal with this problem, many range kernel approximation based acceleration schemes have been proposed. However, all these methods use predefined basis function and assign the same penalisation for all range values to fulfil range kernel approximation. These constraints block from further improving the filtering accuracy and speed. Employing weighted variable projection, a new acceleration method which achieves state-of-the-art performance is proposed. This is done by: (i) unleashing the constraint of using fixed basis function; (ii) exploiting the colour distribution information of the input image to perform a weighted approximation of the range kernel. Experiments demonstrate the superiority of the proposed method in gaining more accurate filtering results efficiently.
References
-
-
1)
-
3. Wang, L., Sheng, Y.: ‘Fast bilateral filtering using exponential functions’, Electron. Lett., 2017, 53, (4), pp. 237–239 (doi: 10.1049/el.2016.4569).
-
2)
-
4. Wang, L., Sheng, Y.: ‘Fast bilateral filtering using the padé approximation’, Electron. Lett., 2017, 53, (6), pp. 395–397 (doi: 10.1049/el.2017.0248).
-
3)
-
5. Hyeong Hong, J., Zach, C., Fitzgibbon, A.: ‘Revisiting the variable projection method for separable nonlinear least squares problems’, The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, USA, July 2017, pp. 5939–5947.
-
4)
-
2. Ghosh, S., Chaudhury, K.N.: ‘On fast bilateral filtering using fourier kernels’, Signal Process. Lett., 2016, 23, (5), pp. 570–573 (doi: 10.1109/LSP.2016.2539982).
-
5)
-
1. Dai, L., Yuan, M., Zhang, X.: ‘Speeding up the bilateral filter: a joint acceleration way’, IEEE Trans. Image Process., 2016, 25, (6), pp. 2657–2672 (doi: 10.1109/TIP.2016.2549701).
http://iet.metastore.ingenta.com/content/journals/10.1049/el.2017.4592
Related content
content/journals/10.1049/el.2017.4592
pub_keyword,iet_inspecKeyword,pub_concept
6
6