The enforcement of social distancing measures has become a critical strategy in public health management to mitigate the spread of contagious diseases. Automated algorithms that estimate interpersonal distances from surveillance footage play a pivotal role in ensuring compliance. Perspective transform is a fundamental technique used to project real-world distances onto image planes, but its computational overhead presents challenges for real-time applications. This paper introduces scalable algorithms for accelerating perspective transform computations using multi-core CPU architectures. By leveraging parallel processing, our approach achieves significant speedups while maintaining accuracy, making it suitable for large-scale deployment in social distancing monitoring systems. Experimental results demonstrate the effectiveness of the proposed methods, with benchmarks showing up-to 2.3 x speedup on an 8-core CPU compared to traditional single-threaded implementations.