A Modified SIFT Descriptor for Image Matching under Spectral Variations
In multispectral imaging multiple discrete wavelength bands are used to image a scene. The imaging process maps the scene contents to different intensity levels and varies the scene appearance from band to band. This induces intensity variations among the spectral images and effects the performance of SIFT for cross spectral image matching. This paper proposes modifications to the SIFT descriptor in order to improve its robustness against spectral variations. The proposed modifications are based on fact, that edges remain well preserved in multispectral imaging and we can achieve better image matching results by boosting the contribution of local edges in the SIFT descriptor construction process. Therefore, we propose a Local Contrast (Δ) and a Differential Excitation (ξ) function for the construction of SIFT descriptors. The experimental results show, that the performance of Δ-SIFT and ξ-SIFT is superior to standard SIFT for image matching under spectral variations.
KeywordsSIFT spectral images interest regions image matching
Unable to display preview. Download preview PDF.
- 1.Brown, M., Su, S.: Multi-spectral SIFT for scene category recognition. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 177–184 (2011)Google Scholar
- 2.Chakrabarti, A., Zickler, T.: Statistics of Real-World Hyperspectral Images. In: IEEE Conference on Computer Vision and Pattern Recognition, pp. 193–200 (2011)Google Scholar
- 4.Hasan, M., Jia, X., Robles-Kelly, A., Zhou, J., Pickering, M.R.: Multi-spectral remote sensing image registration via spatial relationship analysis on SIFT keypoints. In: IEEE International Geoscience and Remote Sensing Symposium, pp. 1011–1014 (2010)Google Scholar
- 5.Ke, Y., Sukthankar, R.: PCA-SIFT: A more distinctive representation for local image descriptors. In: IEEE Conference on Computer Vision and Pattern Recognition, vol. 2, pp. 511–517 (2004)Google Scholar
- 10.Saleem, S., Bais, A., Sablatnig, R.: A performance evaluation of SIFT and SURF for multispectral image matching. In: International Conference on Image Analysis and Recognition, pp. 166–173 (2012)Google Scholar
- 12.Vural, M., Yardimci, Y., Temizel, A.: Registration of multispectral satellite images with orientation-restricted SIFT. IEEE International Geoscience and Remote Sensing Symposium 3, 243–246 (2009)Google Scholar