Optimization of the Pseudocoloring Problem Using Modern Metaheuristics with Hybrid Operators

• Grant number: 25/17551-7
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: November 01, 2025
• End date: October 31, 2026
• Field of knowledge: Physical Sciences and Mathematics - Mathematics - Applied Mathematics
• Principal Investigator: Rodrigo Colnago Contreras
• Grantee: Rafael Damasceno Marques dos Santos
• Host Institution: Instituto de Ciência e Tecnologia (ICT). Universidade Federal de São Paulo (UNIFESP). Campus São José dos Campos. São José dos Campos , SP, Brazil

Abstract

In several situations, specialists need to perform visual analyses of grayscale images. However, the human eye has considerable limitations in identifying details in this type of image, making it necessary to use artificial colorization techniques. The Pseudocoloring Problem (PsCP) consists of assigning to a grayscale image, previously segmented into K subregions, a set of K colors that are as different as possible from each other, in order to facilitate its visual interpretation. This is a problem belonging to the class of NP-hard problems, meaning that there is currently no exact method capable of efficiently solving all its instances within feasible computational time. Thus, the use of metaheuristic-based techniques has proven to be a promising and frequently adopted approach in the literature to address this type of problem. In particular, the Genetic Algorithm (GA) stands out as one of the most widely used techniques in previous approaches to solve the PsCP. This research project aims to propose a new resolution method for the PsCP, based on enhancements of Modern Metaheuristics. The proposal will include the adoption of local search operators and adaptive parameter tuning, implemented through symmetric mapping functions. These mechanisms are expected to contribute to reducing recurring problems in evolutionary algorithms, such as premature convergence and limited exploration of the search space. The proposed method will be evaluated in different scenarios, considering three case studies: (i) the pseudocoloring of real images in the RGB color space; (ii) the pseudocoloring of synthetic and abstract images, where the subregions are completely connected, also in the RGB space; and (iii) the pseudocoloring using the set of colors from the Munsell Color Atlas. Throughout the development of the project, the performance of the proposed method will be compared with other well-established techniques in the literature, aiming to verify its effectiveness and robustness in generating high-quality solutions for the PsCP. It is expected that the methodological enhancements employed-especially the introduction of adaptive operators-will result in a more efficient approach with better generalization capability. Additionally, a library in Python and C/C++ will be developed, implementing the material in a scalable manner for both academic and industrial environments. Finally, the results obtained will be disseminated in internationally recognized journals and conferences.