Wall-Cor Net: wall color replacement via Clifford chance-based deep generative adversarial network

Color design for interior circumstance is a challenging area due to the numerous aspects that must be matched. Although learning from images is a popular method, it is most effective for natural situations with objects that have generally stable colors. The wall color selection is a challenging work because most of the design of the interior depends on a wide range of colors that relate to the location and design of the space. Our objective is to create a system that automatically paints the walls of indoor scene images based on preferred color. To achieve this goal, a novel deep learning-based wall color replacement using deep learning generative adversarial network (WALL-COR NET) has been proposed for the detection of wall color replacement in images. To improve the quality of the input, the indoor scenes are first preprocessed with the Clifford gradient based on RGB representation. Clifford gradient algorithm is employed to analyze the input information and prevent issues like the accumulation of noise distortions in images. To segment the input indoor scene images, a deep learning-based attention V-Net is used to separate the objects from the wall. The GAN helps both designers and homeowners visualize different wall color schemes instantly, saving both time and money. The GAN architecture is utilized for wall color replacement lies in their ability to realistically synthesize diverse color schemes, providing designers and residents to visually explore and choose optimal color schemes for enhanced interior esthetics. Finally, identify suitable and unsuitable images based on user preference, the indoor scene, segmented mask and Hexa color code are supplied to the discriminator and generator. The average classification accuracy for proposed WALL-COR NET is 99.33%. The WALL-COR NET advances the overall accuracy by 19.8%, 17.33% and 4.33, better than FCN, surrogate-assisted method and CNN, respectively.