analiza aiqbench
This research introduces LRConv-NeRV, an optimized variant of Neural Representations for Videos that replaces dense convolutional layers with low-rank separable convolutions to significantly reduce computational cost and memory usage. The study demonstrates that applying this technique selectively to the final decoder stage cuts complexity by 68% while maintaining near-baseline reconstruction quality and achieving a 9.2% bitrate reduction, even under INT8 quantization. These findings offer a practical architectural path for deploying neural video codecs in resource-constrained environments without sacrificing temporal coherence or visual fidelity.
LRConv-NeRV: Low Rank Convolution for Efficient Neural Video Compression
arXiv:2603.18261v2 Announce Type: replace-cross Abstract: Neural Representations for Videos (NeRV) encode entire video sequences within neural network parameters, offering an alternative paradigm to conventional video codecs. However, the convolutional decoder of NeRV remains computationally expensive and memory intensive, limiting its deployment in resource-constrained environments. This paper proposes LRConv-NeRV, an efficient NeRV variant that replaces selected dense 3x3 convolutional layers with structured low-rank separable convolutions, trained end-to-end within the decoder architecture. By progressively applying low-rank factorization from the largest to earlier decoder stages, LRConv-NeRV enables controllable trade-offs between reconstruction quality and efficiency. Extensive experiments demonstrate that applying LRConv only to the final decoder stage reduces decoder complexity by 68%, from 201.9 to 64.9 GFLOPs, and model size by 9.3%, while incurring negligible quality loss and achieving approximately 9.2% bitrate reduction. Under INT8 post-training quantization, LRConv-NeRV preserves reconstruction quality close to the dense NeRV baseline, whereas more aggressive factorization of early decoder stages leads to disproportionate quality degradation. Compared to existing work under layer-aligned settings, LRConv-NeRV achieves a more favorable efficiency versus quality trade-off, offering substantial GFLOPs and parameter reductions while maintaining higher PSNR/MS-SSIM and improved temporal stability. Temporal flicker analysis using LPIPS further shows that the proposed solution preserves temporal coherence close to the NeRV baseline, results establish LRConv-NeRV as a potential architectural alternative for efficient neural video decoding under low-precision and resource-constrained settings.
Abstract:Neural Representations for Videos (NeRV) encode entire video sequences within neural network parameters, offering an alternative paradigm to conventional video codecs. However, the convolutional decoder of NeRV remains computationally expensive and memory intensive, limiting its deployment in resource-constrained environments. This paper proposes LRConv-NeRV, an efficient NeRV variant that replaces selected dense 3x3 convolutional layers with structured low-rank separable convolutions, trained end-to-end within the decoder architecture. By progressively applying low-rank factorization from the largest to earlier decoder stages, LRConv-NeRV enables controllable trade-offs between reconstruction quality and efficiency. Extensive experiments demonstrate that applying LRConv only to the final decoder stage reduces decoder complexity by 68%, from 201.9 to 64.9 GFLOPs, and model size by 9.3%, while incurring negligible quality loss and achieving approximately 9.2% bitrate reduction. Under INT8 post-training quantization, LRConv-NeRV preserves reconstruction quality close to the dense NeRV baseline, whereas more aggressive factorization of early decoder stages leads to disproportionate quality degradation. Compared to existing work under layer-aligned settings, LRConv-NeRV achieves a more favorable efficiency versus quality trade-off, offering substantial GFLOPs and parameter reductions while maintaining higher PSNR/MS-SSIM and improved temporal stability. Temporal flicker analysis using LPIPS further shows that the proposed solution preserves temporal coherence close to the NeRV baseline, results establish LRConv-NeRV as a potential architectural alternative for efficient neural video decoding under low-precision and resource-constrained settings.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2603.18261 [cs.CV]
(or arXiv:2603.18261v2 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2603.18261 Focus to learn more arXiv-issued DOI via DataCite
IEEE Access Vol. 14 pp. 97418 to 97431 in 2026
0.1109/ACCESS.2026.3707284 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 336 cuvinte
Subiecte conexe
[AINews] not much happened today
Latent Space · 10 sept. 2026
[AINews] High Taste Testers Ep. 1
Latent Space · 17 oct. 2026
[AINews] not much happened today
Latent Space · 3 oct. 2026
Group-Shared Low-Rank Approximation for Mobile-Efficient Pointwise Convolutions in Large-Kernel CNNs
arXiv · 27 aug. 2026