A proposed cooperative approach for edge detection based multi-agent system (MAS) using heterogeneous RGB images

Image segmentation has become a growing research area in image processing, computer vision and pattern recognition. Despite decades of progress and significant accomplishments, challenges remain in feature extraction and model design for image segmentation which make it as complex system required to be solved by intelligent system. This manuscript systematically proposes a cooperative approach for images edge detection based on Multi-Agent System, the proposed approach advance addresses ongoing obstacles of existing image processing methods. Regarding to the image segmentation methods issues is extracting features and providing accurate segmented images. Often edges and regions belonging to the input image are confusing and intricate to extract, due to the proximity of the brightness levels of the images. Many technologies and approaches have been proposed for the identification, classifications, detection and segmentation of images. However, the optimal solution remains a challenging task in computer vision. Our proposed Multi-Agent System(MAS) framework adopts a bottom-up approach to enhance the Quadtree Algorithm while also leveraging a bottom-up strategy through a fusion process facilitated by dynamic aggregation process. This cooperative protocol among region and edge agents enables the framework to achieve more accurate and precise segmentation compared to state-of-the-art methods, even when applied to heterogeneous images from diverse benchmarks. Our approach out performs the compared state of the art techniques with PSNR of 19.2107 dB, MSE of 0.0059, and SSIM of 0.9602.