Towards Automatic Image Enhancement with Genetic Programming and Machine Learning

Image Enhancement (IE) is an image processing procedure in which the image’s original information improved, highlighting specific features to ease post-processing analyses by a human or machine. State-of-the-art enhancement pipelines apply solutions fixed and static constraints solve issues isolation. In this work, IE system for marketing proposed, more precisely, real estate marketing, where objective enhance commercial appeal of images, while maintaining level realism similarity with image. This work proposes generic pipeline that combines state-of-the-art filters, Machine Learning methods, Evolutionary approaches, such as Genetic Programming (GP), create dynamic framework Enhancement. The GP-based trained optimize 4 metrics: Neural Assessment (NIMA) technical BRISQUE, evaluate quality images; NIMA aesthetics PhotoILike, attractiveness. It shown GP model was able find best (0.97 Aesthetics), high images (Structural Similarity Index Measure (SSIM) 0.88). has better performance according metrics than off-the-shelf tool framework’s isolated parts.