Physics Maths Engineering
Seyed Hossein Amirshahi,
Seyed Hossein Amirshahi
Amirkabir University of Technology (Tehran Polytechnic), School of Material Engineering and Advanced Processes
Ida Rezaei,
Ida Rezaei
Amirkabir University of Technology (Tehran Polytechnic), School of Material Engineering and Advanced Processes
Ali Akbar Mahbadi
Ali Akbar Mahbadi
Amirkabir University of Technology (Tehran Polytechnic), School of Material Engineering and Advanced Processes
Peer Reviewed
Two regression methods, namely, Support Vector Regression (SVR) and Kernel Ridge Regression (KRR), are used to reconstruct the spectral reflectance curves of samples of Munsell dataset from the corresponding CIE XYZ tristimulus values. To this end, half of the samples (i.e., the odd ones) were used as training set while the even samples left out for the evaluation of reconstruction performances. Results were reviewed and compared with those obtained from Principal Component Analysis (PCA) method, as the most common context-based approach. The root mean squared error (RMSE), goodness fit coefficient (GFC), and CIE LAB color difference values between the actual and reconstruct spectra were reported as evaluation metrics. However, while both SVR and KRR methodologies provided better spectral and colorimetric performances than the classical PCA method, the computation costs were considerably longer than PCA method.
Spectral reflectance reconstruction is the process of predicting the full spectral reflectance curve of a material (how it reflects light at different wavelengths) from limited color data, such as CIE XYZ tristimulus values. This is important for accurate color reproduction in industries like printing, textiles, and digital imaging.
The study used two advanced regression methods: Support Vector Regression (SVR) and Kernel Ridge Regression (KRR). These were compared with the traditional Principal Component Analysis (PCA) method to see which performs best in reconstructing spectral reflectance curves.
The Munsell dataset, a standard collection of color samples, was split into two parts: odd-numbered samples were used for training the models, and even-numbered samples were used to test the accuracy of the reconstructed spectra.
Both SVR and KRR outperformed PCA in terms of spectral and colorimetric accuracy. They achieved lower Root Mean Squared Error (RMSE), higher Goodness Fit Coefficient (GFC), and smaller CIE LAB color differences compared to PCA.
While SVR and KRR provided better results, they required significantly more computation time compared to PCA. This makes them less efficient for large-scale applications where speed is critical.
PCA is a widely used method for dimensionality reduction and spectral reconstruction. It served as a baseline to compare the performance of SVR and KRR, highlighting the trade-offs between accuracy and computational cost.
The study used three key metrics:
Accurate spectral reflectance reconstruction ensures precise color reproduction in applications like digital imaging, printing, and material design. It helps maintain color consistency across different devices and lighting conditions.
Yes, SVR and KRR can be used in industries requiring high color accuracy, such as graphic design, textile manufacturing, and digital displays. However, their higher computational cost may limit their use in time-sensitive applications.
Future research could focus on optimizing SVR and KRR for faster computation or combining them with other techniques to balance accuracy and efficiency. Expanding the dataset and testing on real-world color samples could also improve their practicality.
Show by month | Manuscript | Video Summary |
---|---|---|
2025 April | 3 | 3 |
2025 March | 74 | 74 |
2025 February | 45 | 45 |
2025 January | 50 | 50 |
2024 December | 42 | 42 |
2024 November | 41 | 41 |
2024 October | 36 | 36 |
2024 September | 57 | 57 |
2024 August | 36 | 36 |
2024 July | 33 | 33 |
2024 June | 21 | 21 |
2024 May | 29 | 29 |
2024 April | 23 | 23 |
2024 March | 6 | 6 |
Total | 496 | 496 |
Show by month | Manuscript | Video Summary |
---|---|---|
2025 April | 3 | 3 |
2025 March | 74 | 74 |
2025 February | 45 | 45 |
2025 January | 50 | 50 |
2024 December | 42 | 42 |
2024 November | 41 | 41 |
2024 October | 36 | 36 |
2024 September | 57 | 57 |
2024 August | 36 | 36 |
2024 July | 33 | 33 |
2024 June | 21 | 21 |
2024 May | 29 | 29 |
2024 April | 23 | 23 |
2024 March | 6 | 6 |
Total | 496 | 496 |