Cross-Racial Face Detection Performance Analysis: A Comparative Study of YCrCb and YCrCb Color Spaces under Varied Environmental Conditions

This paper presents a robust comparative framework for face detection under varying environmental conditions through chromatic analysis in $Y C_{r} C_{b}$ and $Y C_{r}^{\prime} C_{b}^{\prime}$ color spaces, demonstrating geometric invariance to translation, rotation, and scale transformations. The methodology initiates with color space conversion of input facial images, followed by adaptive thresholding techniques for skin segmentation—employing optimized $C_{b} C_{r}$ and enhanced $C_{b}^{\prime} C_{r}^{\prime}$ channel boundaries derived from empirical skin distribution analysis. Subsequent refinement stages incorporate: (i) Morphological filtering using elliptical structuring elements for noise suppression, (ii) Geometric feature selection based on aspect ratio and compactness criteria, (iii) Multi-scale validation through pyramid decomposition Experimental results quantify detection accuracy improvements when employing the enhanced $Y C_{r}^{\prime} C_{b}^{\prime}$ model compared to conventional $Y C_{r} C_{b}$ approaches. The proposed dual-space analysis establishes a theoretical foundation for adaptive color space selection in cross-demographic face detection systems.