Image Region Selection for Color Reproducibility

The patent introduces three distinct evaluation indices (evenness degree, color purity, spatial dispersion degree) that assess different local qualities of segment regions. Each index captures a specific aspect of region suitability: evenness degree measures density uniformity, color purity measures proximity to color materials, and spatial dispersion degree measures distribution across the image. By combining these multiple local quality assessments, the system achieves accurate color reproducibility evaluation without excessive processing complexity.

The patent divides the image into multiple segment regions and extracts several candidate object regions that satisfy predetermined conditions for evenness degree, color purity, and spatial dispersion degree. By evaluating multiple regions rather than relying on a single region, the system ensures that color reproducibility adjustment is based on comprehensive information from different parts of the image, thereby maintaining consistent color reproduction across the entire image while managing processing time through efficient region selection.

The patent establishes predetermined threshold conditions for three key parameters (evenness degree, color purity, spatial dispersion degree) to filter and select candidate object regions. By transforming the complex optimization problem into a parameter-based filtering approach, the system efficiently identifies suitable segment regions without exhaustively evaluating all possible combinations. This parameter change strategy maintains high color reproducibility evaluation accuracy while significantly reducing processing time compared to brute-force methods.