G. Van de Wouwer, P. Scheunders, S. Livens, and D. Van Dyck. Colour texture classification by wavelet energy-correlation signatures. Pattern Recognition, 1997. to appear. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Non-Linear Mapping for Feature Extraction - Scheunders, De Backer, Naud (1998) Self-citation (Scheunders) (Correct)
....set is estimated to be about 3. 3. A colour texture data base of N = 1024 images containing 16 different textures (16 classes) The feature vectors are generated by employing a discrete wavelet transform on the R, G and B plane separately and calculating the energy features of the detail images [14]. A 48 dimensional data space is obtained. The intrinsic dimensionality is estimated to be about 5. In the conducted experiments, the SOM and AFN are compared to linear dimensionality reduction. For this purpose, a Principal Component Analysis (PCA) is performed. All the algorithms are written in ....
G. Van de Wouwer, P. Scheunders, S. Livens, and D. Van Dyck. Colour texture classification by wavelet energy-correlation signatures. Pattern Recognition, 1997. to appear.
Wavelet-based Texture Analysis - Scheunders, Livens, Wouwer, Vautrot, .. (1997) (3 citations) Self-citation (Wouwer Scheunders Livens Van dyck) (Correct)
....planes. Features wwith equal variances are obtained by the correlation elements: C X j X k ni = C X j X k ni q C X j X j ni q C X k X k ni (17) A recent study by Van de Wouwer et al. reveals that using these colour features offers substantially more discriminative information [21]. It is also shown that the choice of the colour space can lead to significant differences in performance. Jain and Healey use opponent features computed from Gabor filter outputs and combine information across spectral bands at different scales [22] Remark that the same arguments hold for ....
....set, for all four cases. From the results it is clear that the use of the correlation features improves classification results. Transformation to other colour spaces improves classification performance. A detailed description of this experiment and a discussion on the results can be found in [21]. C. Experiment 3 In the last experiment, rotation invariant texture classification is the issue. We will demonstrate the usefulness of the second order spectral moments. 10 natural textured images from the Brodatz album which clearly exhibit anisotropy were chosen and are displayed in Fig. 5. ....
G. Van de Wouwer, P. Scheunders, S. Livens, and D. Van Dyck, "Colour texture classification by wavelet energy-correlation signatures," Pattern Recognition, 1997, to appear.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC