Iterative Graph-Based HDR Image Enhancement
Zachary McBride Lazri, Guan‐Ming Su · 2023
Digital image enhancement has long been an active research area in the field of image processing. While many developed techniques apply global enhancements to an entire image or local enhancements in pixel neighborhoods, advanced segmentation tools provide the opportunity for more flexible object-specific processing. In this paper, we focus on improving the visual quality within an object while also improving its quality with respect to its neighbors. Objects are represented as point clouds in luminance-saturation space and are separated from each other while maintaining an image's natural integrity using an iterative graph-based approach. The nodes in the graph represent objects in an image, and in each iteration, inter-object enhancement is applied to make neighboring objects stand out with respect to each other. Intra-object enhancement is also applied in each iteration to improve the contrast and details within each object. Experimental results demonstrate the effectiveness of this algorithm in enhancing an image's quality.