Wiener estimation is a widely used technique for the spectral reflectance reconstruction of colored objects. In our study, the reflectance estimation of 3-channel images acquired using a traditional RGB camera was compared with that of 6-channel images captured with a modified 6-channel camera. 240 patches of ColorChecker DC Chart were used for training and testing of the models. Their performance was assessed via the quality metrics PSNR, RMSE and CIEDE2000 based on the reconstructed and spectrophotometrically measured reflectance values. Compared to 3-channel models, 6-channel models in general provided better results. The lowest color difference – CIEDE2000 – was obtained when using 10 calculation terms in the 3-channel algorithm and 13 terms in the 6-channel. The best reconstruction models, i.e. the lowest CIEDE2000 and RMSE, and the highest PSNR values, were found for the skin and neutral tones while the performance was the poorest with the saturated color patches. The obtained results can be of use in various practical applications, e.g. in a printing workflow with high precision color rendering or for accurate digitization of paintings and other artworks.
Due to the complexity of individual steps within the multi-channel workflow (from capture to print) and understanding of the challenges and specificities of each step, the dissertation focused solely on the capture process. Consequently, during the research work which took place in Darmstadt, we successfully captured not only test charts but also artworks using the modified camera, thereby effectively digitizing and preserving them for further use. This format still holds a significant potential and provides a basis for engaging research and further analysis using various algorithms in the future.
In the second part of the research, we also focused on the colorimetric accuracy of conventional digital cameras, which was evaluated based on the CIEDE2000 metrics and the creation of camera ICC profiles, as well as the calculated profile volumes. By comparing the average CIEDE2000 color differences between the cameras and profile volumes, we observed a trend indicating that cameras with larger profile volumes (color gamut) provide lower CIEDE2000 color difference values, and vice versa.
Based on the results and comparison between the capture using 3- and 6-channel devices, it can be concluded that the use of a 6-channel camera resulted in a highly accurate (in terms of colorimetry) multi-channel capture, as indicated by the average CIEDE2000 differences, which showed a significant improvement in favor of the 6-channel capture.
|
