Deep Joint Source-Channel Coding for DNA Image Storage: A Novel Approach With Enhanced Error Resilience and Biological Constraint Optimization

Abstract

In the current era, DeoxyriboNucleic Acid (DNA) based data storage emerges as an intriguing approach, garnering substantial academic interest and investigation. This paper introduces a novel deep joint source-channel coding (DJSCC) scheme for DNA image storage, designated as DJSCC-DNA. This paradigm distinguishes itself from conventional DNA storage techniques through three key modifications: 1) it employs advanced deep learning methodologies, employing convolutional neural networks for DNA encoding and decoding processes; 2) it seamlessly integrates DNA polymerase chain reaction (PCR) amplification into the network architecture, thereby augmenting data recovery precision; and 3) it restructures the loss function by targeting biological constraints for optimization. The performance of the proposed model is demonstrated via numerical results from specific channel testing, suggesting that it surpasses conventional deep learning methodologies in terms of peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). Additionally, the model effectively ensures positive constraints on both homopolymer run-length and GC content.