Entropy-driven exposure interpolation for large exposure-ratio imagery

Sensor limitations in capturing devices and environmental factors can result in radiance artifacts in rendered images. This paper presents an entropy-driven exposure interpolation framework in the context of large exposure-ratio fusion. The proposed framework generates intermediate exposure-corrected images through transmission map estimation to obtain initial radiance and illumination maps. Fusion weight maps, within a pyramidal framework, are derived from the transmission map and spatial entropy, thereby enhancing the visual quality of images while preventing color artifacts. Experiments demonstrate that the proposed framework outperforms several state-of-the-art multi-exposure fusion schemes.