Applying Spectral Decomposition Theory To Spectral Data of Textile Dyed Samples For Decrease Color Difference Between Actual and Reconstructed Data Using Pca

Massive volumes of spectral data require compression to facilitate spectral imaging transmission and storage. the principal component analysis is a classic compression method. this study carried out pre-analyzed data with the aid of spectral decomposition theory before applying the principal component analysis operation. attempts were made to compress a series of diversely scattered spectral data of dyed samples using fundamental color stimuli (rfcs). the results show that the reconstruction from rfcs was ideal, giving zero color difference under the illuminant/observer combination used for the rfcs formation. also, the results show that although the color coordinates of the dyed textile samples are diversely scattered in the chosen color space, the compression through fundamental color stimuli gives far better results than compression through measured reflectance data.

Applying Spectral Decomposition Theory to Spectral Data of Textile dyed Samples for decrease Color Difference between Actual and Reconstructed Data using PCA

Massive volumes of spectral data require compression to facilitate spectral imaging transmission and storage. The principal component analysis is a classic compression method. This study carried out pre-analyzed data with the aid of spectral decomposition theory before applying the principal component analysis operation. Attempts were made to compress a series of diversely scattered spectral data of dyed samples using fundamental color stimuli (RFCS). The results show that the reconstruction from RFCS was ideal, giving zero color difference under the illuminant/observer combination used for the RFCS formation. Also, the results show that although the color coordinates of the dyed textile samples are diversely scattered in the chosen color space, the compression through fundamental color stimuli gives far better results than compression through measured reflectance data.