Skip to Main content Skip to Navigation
Journal articles

Design of estimators for restoration of images degraded by haze using genetic programming

Abstract : Restoring hazy images is challenging since it must account for several physical factors that are related to the image formation process. Existing analytical methods can only provide partial solutions because they rely on assumptions that may not be valid in practice. This research presents an effective method for restoring hazy images based on genetic programming. Using basic mathematical operators several computer programs that estimate the medium transmission function of hazy scenes are automatically evolved. Afterwards, image restoration is performed using the estimated transmission function in a physics-based restoration model. The proposed estimators are optimized with respect to the mean-absolute-error. Thus, the effects of haze are effectively removed while minimizing overprocessing artifacts. The performance of the evolved GP estimators given in terms of objective metrics and a subjective visual criterion, is evaluated on synthetic and real-life hazy images. Comparisons are carried out with state-of-the-art methods, showing that the evolved estimators can outperform these methods without incurring a loss in efficiency, and in most scenarios achieving improved performance that is statistically significant.
Complete list of metadata
Contributor : Pierrick Legrand Connect in order to contact the contributor
Submitted on : Tuesday, October 30, 2018 - 6:44:11 PM
Last modification on : Thursday, January 20, 2022 - 5:31:37 PM


  • HAL Id : hal-01909121, version 1



José Enrique Hernandez-Beltran, Victor H. Díaz-Ramírez, Leonardo Trujillo, Pierrick Legrand. Design of estimators for restoration of images degraded by haze using genetic programming. Swarm and Evolutionary Computation, Elsevier, 2018. ⟨hal-01909121⟩



Les métriques sont temporairement indisponibles