A Hardware Architecture Based on Genetic Clustering for Color Image Segmentation

Color image segmentation finds several real-life applications on hyperspectral image processing, brain tumor detection (Biomedical), facial recognition (Biometric), object tracking (Video analysis), etc. In this manuscript, the color image segmentation is dealt as a clustering problem. A genetic algorithm (GA)-based hardware architecture is proposed to perform the segmentation task in a fast manner. Testing of the proposed architecture is carried out on four standard RGB color images like Pepper, Baboon, Lenna, and Colorbars. Comparison with three other benchmark architectures of genetic algorithm reveals that the proposed architecture provides satisfactory results in terms of complexity, system clock frequency, and resource utilization. The three other architectures used for comparison are compact implementation of GA, used for simple optimization tasks, whereas the proposed one is used for clustering huge number of pixels within an image, for executing the task of segmentation.