DNA-based storage has emerged as a promising approach to the global data crisis, offering molecular-scale density and millennial-scale stability at low maintenance cost. Over the past decade, substantial progress has been made in storing text, images, and files in DNA -- yet video remains an open challenge. The difficulty is not merely technical: effective video DNA storage requires co-designing compression and molecular encoding from the ground up, a challenge that sits at the intersection of two fields that have largely evolved independently. In this work, we present HELIX, the first end-to-end neural network jointly optimizing video compression and DNA encoding -- prior approaches treat the two stages independently, leaving biochemical constraints and compression objectives fundamentally misaligned. Our key insight: token-based representations naturally align with DNA's quaternary alphabet -- discrete semantic units map directly to ATCG bases. We introduce TK-SCONE (Token-Kronecker Structured Constraint-Optimized Neural Encoding), which achieves 1.91 bits per nucleotide through Kronecker-structured mixing that breaks spatial correlations and FSM-based mapping that guarantees biochemical constraints. Unlike two-stage approaches, HELIX learns token distributions simultaneously optimized for visual quality, prediction under masking, and DNA synthesis efficiency. This work demonstrates for the first time that learned compression and molecular storage converge naturally at token representations -- suggesting a new paradigm where neural video codecs are designed for biological substrates from the ground up.
Figures & tables
Figure 1: HELIX: End-to-end pixel-to-nucleotide pipeline. (a) System architecture with dual-stream tokenization, TK-SCONE encoding, DNA synthesis/sequencing, and transformer-based reconstruction. Inset shows DNA strand structure with primers, address index, encoded payload, and Reed-Solomon parity. (b) TK-SCONE internal architecture showing two-stage processing: Kronecker adaptive mixing for correlation breaking followed by FSM-based quaternary mapping for constraint compliance.
Figure 2: Kronecker mixing effect. Direct binary-to-DNA mapping (left) violates biochemical constraints with long homopolymers and GC imbalance. Kronecker transform (center) redistributes correlations through GF(2) mixing. Result (right): balanced sequences achieving 1.91 bpn with minimal padding.
Figure 3: Cost-quality Pareto frontier.
Figure 4: Visual comparison at different mask rates, shown in the format mask rate @ LPIPS (lower is better).
Method
bpn↑
$/frame↓
LPIPS↓
PSNR↑
Token+Goldman [ 11 ]
1.58
11.06
0.29
28.4
Token+Press [ 8 ]
1.83
9.59
0.29
28.4
Ours (no mask)
1.91
9.10
0.27
29.2
Ours (60% mask)
1.91
3.64
0.28
28.8
Ours (70% mask)
1.91
2.73
0.30
28.1
Table 1: End-to-end vs. two-stage approaches on UVG.
Configuration
bpn
$/frame
LPIPS
Δ LPIPS
Full system
1.91
3.64
0.280
–
w/o Kronecker
1.68
4.15
0.283
+0.003
w/o masking
1.91
9.10
0.270
-0.010
w/o transformer
1.91
3.64
0.352
+0.072
Table 2: Component ablation at 60% mask rate on UVG.
While neural lossless image compression has advanced significantly with learned entropy models, lossless video compression remains largely unexplored in the neural setting. We present NeuralLVC, a neural lossless video codec that combines masked diffusion with an I/P-frame architecture for exploiting temporal redundancy. Our I-frame model compresses individual frames using bijective linear tokenization that guarantees exact pixel reconstruction. The P-frame model compresses temporal differences between consecutive frames, conditioned on the previous decoded frame via a lightweight reference embedding that adds only 1.3% trainable parameters. Group-wise decoding enables controllable speed-compression trade-offs. Our codec is lossless in the input domain: for video, it reconstructs YUV420 planes exactly; for image evaluation, RGB channels are reconstructed exactly. Experiments on 9 Xiph CIF sequences and on 24 HD sequences at 720p and 1080p (Xiph, UVG, and the modern BVI-AOM benchmark) show that NeuralLVC outperforms H.264 and H.265 lossless by a significant margin at every resolution tested. We verify exact reconstruction through end-to-end encode-decode testing with arithmetic coding. These results suggest that masked diffusion with temporal conditioning is a promising direction for neural lossless video compression.
Tiberio Uricchio, Marco Bertini
University of Pisa · Pisa, Italy · Università degli Studi di Firenze +1
Video large language models (Video-LLMs) represent videos as dense sequences of visual tokens, whose length grows with the temporal and spatial extent of the input. These tokens often contain substantial redundancy arising from repeated visual patterns, leading to unnecessary computation in the subsequent language-model processing. Existing token compression methods, including pruning and merging, perform compression online during inference, repeatedly incurring additional computation for each input video and often relying on model-specific designs that limit their generality, we instead rethink this paradigm by shifting the costly compression process offline. We propose ONCE, a plug-in video token compression framework that introduces an offline-to-online paradigm: a frequency-aware global codebook is learned once in the visual feature space and reused for lightweight online compression through codebook lookup and aggregation, reducing repeated per-video computation and the need for model-specific compression designs. Extensive experiments across multiple video understanding benchmarks and against diverse compression baselines demonstrate that our approach achieves a strong accuracy-efficiency trade-off, maintaining competitive performance while achieving the lowest inference latency among compared methods.
Jiayang He, Tianling Xu, Diancheng Kang +4
Southern University of Science and Technology · Imperial College London · Jilin University
Video language models process videos as dense visual-token sequences with substantial representational redundancy. Compressing these sequences is therefore essential for reducing the visual-token burden on language-model decoding. The central challenge is to preserve visual information dispersed across frames under such compression. To this end, we introduce Aggregating Visual Information with Optimal Transport (AVIOT), which casts video token compression as transporting a dense empirical measure of frame observations onto a compact target measure. The resulting source-to-target coupling induces a distribution over source observations for each target support, directly specifying how the compressed video representation is constructed. We further adapt this construction along task and spatial axes. Question conditioning modulates the transport cost between source frames and target supports, while influencing how many supports are allocated to each temporal segment, thereby directing representation capacity toward question-relevant content. At multiple spatial granularities, AVIOT computes region-specific temporal transport plans and adaptively fuses the representations they yield, allowing different regions within the same compact representation to draw from different moments. Evaluations across varying compression ratios show that AVIOT matches or outperforms the uncompressed baseline on multiple video-understanding benchmarks while retaining strong performance at higher compression ratios.
Wenti Yin, Xiaotian Han, Junyuan Shang +5
Key Laboratory of Image Processing and Intelligent Control, School of Artificial Intelligence and Automation, Huazhong University of Science and Technology · Baidu, Inc