High Fault-Tolerant DNA Image Storage System Based on VAE.

Area of Science:

• Biotechnology
• Data Storage
• Computer Science

Background:

• DNA-based storage offers vast potential but faces challenges due to error-prone synthesis and sequencing.
• Image data compression for DNA storage is vulnerable to errors, potentially causing data loss during decompression.

Purpose of the Study:

• To develop a high fault-tolerant DNA image storage system to improve compression ratio and data resilience.
• To address the limitations of current DNA storage methods for image data.

Main Methods:

• Image compression using Variational Autoencoder (VAE) and Singular Value Decomposition (SVD).
• Quantization and rotational coding of latent variables to ensure DNA sequence constraints (homopolymer length, GC content).
• Optimized error-correction schemes tailored to specific error types for improved data recovery.

Main Results:

• Demonstrated significant improvements in image data compression ratio and fault tolerance.
• Achieved superior resilience against errors compared to existing methods.
• Validated the system's performance through image compression simulations and adjustable compression ratios via image scaling.

Conclusions:

• The proposed system enhances the reliability and efficiency of DNA-based image storage.
• The innovative combination of VAE, SVD, and optimized error correction overcomes key challenges in DNA data storage.
• This approach paves the way for practical applications of DNA storage for high-density image archiving.