AI-Driven Framework for Medical Image Compression and security Using Deep Learning

Vamsi¹, K.Hari krishna²

¹Assistant Professor, ²MCA Final Semester, Master of Computer Applications,

Sanketika Vidya Parishad Engineering College, Visakhapatnam, Andhra Pradesh, gvkrishna2002@gmail.com,harikavadana@gmail.com

ABSTRACT

The rapid growth of medical imaging data from modalities such as CT, MRI, X-ray, and ultrasound poses significant challenges in efficient storage, secure transmission, and patient privacy protection, particularly in telemedicine and cloud-based healthcare systems. Conventional image compression methods often compromise diagnostic quality, while traditional encryption techniques limit the direct use of artificial intelligence (AI) on protected data. This work presents a simple and practical AI-driven framework for medical image compression and security using a lightweight convolutional auto encoder combined with AES-256 encryption. The auto encoder compresses medical images into a compact latent representation while preserving clinically important information.AES-256 encryption is applied to the compressed data to ensure strong confidentiality during storage and transmission. At the receiver side, the encrypted data is decrypted and accurate l reconstructed using the decoder network. Experimental results on benchmark medical image datasets demonstrate effective compression performance and high quality image reconstruction, with minimal computational overhead introduced by encryption. The proposed framework is well suited for secure and efficient real-world medical image applications.

Keywords: Medical Image Compression, Convolutional Auto encoder, AES-256 Encryption, Deep Learning, Healthcare Data Security, Privacy Preservation