| Aseem Agarwala, Mira Dontcheva, Maneesh Agrawala, Steven Drucker, Brian Curless, David Salesin, Michael Cohen
This paper presents an interactive, computer-assisted framework for digital photomontage, a process of combining parts of multiple photographs into a single composite image. The framework uses graph-cut optimization to select seams between images for seamless integration and gradient-domain fusion to reduce visible artifacts. It also includes interactive tools that allow users to specify high-level image objectives, either globally or locally, to guide the composite creation process. The framework is versatile and can be applied to various tasks such as selective composites, extended depth of field, relighting, panoramic stitching, clean-plate production, stroboscopic visualization of movement, and time-lapse mosaics. The system allows users to iteratively refine the composite by applying different objectives, and then use gradient-domain fusion to improve the final result. The paper also discusses related work, the technical aspects of the framework, and presents results demonstrating the effectiveness of the approach in various applications. The framework is designed to be user-friendly and efficient, enabling the creation of high-quality composites with minimal manual intervention.This paper presents an interactive, computer-assisted framework for digital photomontage, a process of combining parts of multiple photographs into a single composite image. The framework uses graph-cut optimization to select seams between images for seamless integration and gradient-domain fusion to reduce visible artifacts. It also includes interactive tools that allow users to specify high-level image objectives, either globally or locally, to guide the composite creation process. The framework is versatile and can be applied to various tasks such as selective composites, extended depth of field, relighting, panoramic stitching, clean-plate production, stroboscopic visualization of movement, and time-lapse mosaics. The system allows users to iteratively refine the composite by applying different objectives, and then use gradient-domain fusion to improve the final result. The paper also discusses related work, the technical aspects of the framework, and presents results demonstrating the effectiveness of the approach in various applications. The framework is designed to be user-friendly and efficient, enabling the creation of high-quality composites with minimal manual intervention.