The deployment of Large Intelligent Surfaces (LIS) has demonstrated considerable promise in improving the performance of wireless networks. However, the implementation of such systems is often complicated by their substantial architectural complexity. In this research, a refined strategy is presented wherein a limited array of antenna elements is selectively utilized to reduce the system's operational demands. Experimental evaluations demonstrate that progressively scaling the number of active units attenuates efficiency degradation, thereby facilitating an optimal trade-off between computational performance and resource utilization. When integrated with Low-Density Parity-Check (LDPC) error correction schemes, this selective approach offers a scalable and pragmatic alternative to full-scale LIS configurations, ideally suited for complex communication networks.