Design of an Embedded Image Compression System Using Python and Machine Learning

The exponential growth of digital imagery necessitates advanced compression techniques that balance storage efficiency, transmission speed, and image quality. This paper presents an embedded image compression system leveraging Python and machine learning, integrating a convolutional neural network (CNN) for feature extraction and a variational autoencoder (VAE) for compression. The system, implemented on a Raspberry Pi 4 Model B, achieves an average compression ratio of 20:1, with a maximum of 25:1, while maintaining high image fidelity, as evidenced by an average PSNR of 32 dB and SSIM of 0.90. The real-time processing capability allows compression at 20 images per second, making it suitable for resource-constrained environments. Comparative analysis with JPEG and JPEG 2000 highlights the system's superior compression efficiency and image quality. Benchmarking on high-performance computing setups further validates its scalability, achieving a compression ratio of 22:1 with a PSNR of 33 dB. These results underscore the transformative role of machine learning in embedded image compression, paving the way for more efficient, real-time applications in digital imaging. This system is especially suitable for sustainable healthcare applications such as X-ray transmission and telemedicine imaging, where efficient bandwidth usage and image fidelity are critical for remote diagnosis and storage scalability.