CARA (Collision-Aware Resolution Adaptation) is a technique that dynamically assigns resolution levels in multiresolution hash encodings to balance information load across hash tables, thereby reducing hash collisions and improving parameter efficiency. It also introduces an invertible pixel‑shuffle transform to redistribute spatial information and lower hash load factors without enlarging tables. Experiments on Kodak images, gigapixel natural images, and a newly curated uncompressed whole‑slide image dataset show that CARA consistently enhances the fidelity‑parameter trade‑off, matching state‑of‑the‑art performance with only 27.76% of the parameters and achieving up to 6.11 dB PSNR improvement at comparable parameter counts.
Diffusion-based generative models have achieved remarkable success in real-world image super-resolution (SR). With tiled diffusion techniques, these models can produce high-resolution images that exceed their native-supported resolution.
CC-4DGS introduces a storage‑efficient framework for dynamic 4D Gaussian splatting by replacing large multi‑resolution hash tables with a deterministic dense hash encoding and compact neural decoders, reducing deformation storage to 1–3 MB per scene. It also compresses canonical point‑cloud attributes through conditional autoencoding, selective quantization, and residual codebooks, achieving 3–5× reduction in point‑cloud size with negligible quality loss. The combined approach maintains real‑time rendering performance while lowering total storage to 20–30 MB, matching state‑of‑the‑art reconstruction accuracy on N3DV and Technicolor Light Field datasets.
arXiv:2608.30782v1 Announce Type: new Abstract: Real-world image super-resolution (Real-ISR) aims to preserve structures supported by the degraded observation while reconstructing perceptually realis...
arXiv:2410.13925v2 Announce Type: replace Abstract: \textit{Nature is infinitely resolution-free}. In the context of this reality, existing diffusion models, such as Diffusion Transformers, often fac...
BudgetPix introduces a compute‑adaptive tokenization framework for pixel‑space image diffusion, dynamically allocating computational resources based on visual complexity and spatial layout. It comprises an adaptive encoder using an entropy‑guided quadtree and multi‑scale patch embedder, a scale‑aware decoder, and a flexible training schedule that allows pixel‑space denoisers to operate across variable token counts. The method integrates with existing diffusion architectures, achieving comparable fidelity to larger models while using only a fraction of the compute budget, and demonstrates improved quality‑efficiency tradeoffs in both text‑to‑image and class‑conditional generation.
GigaBench is the first benchmark designed to evaluate learnable representations (LRs) for gigapixel image compression and super‑resolution. It compiles a diverse set of gigapixel images—collected, acquired, and synthetically rendered—from various domains and tests representative LR methods (Implicit, Hybrid, and Explicit) on consumer hardware. The study systematically measures reconstruction quality, model size, compression, and inference times, revealing consistent trends and offering actionable guidance for future research.
arXiv:2606. 27978v1 Announce Type: cross Abstract: Pixel-space continuous-token autoregressive (AR) generation directly models images as sequences of raw pixel patches, avoiding discrete tokenization or a separately pretrained tokenizer.
arXiv:2606. 29400v1 Announce Type: cross Abstract: In computer graphics, visual content is continuously warped, zoomed and resampled.
MoE-ViE introduces a Mixture-of-Experts vision encoder that scales efficiently for image and video understanding, outperforming dense counterparts across various sizes. The study shows fine‑grained MoE topologies provide significant gains, and proposes an auxiliary‑loss‑free balancing variant and a specialized MoE kernel to reduce inference latency. With frame‑level distillation and a novel freezing mechanism, the largest MoE‑ViE model matches state‑of‑the‑art zero‑shot performance while being 1.7× larger and 76% faster, and it outperforms other encoders when paired with a language model on both image and video benchmarks.
arXiv:2508.15774v2 Announce Type: replace Abstract: Visual diffusion models achieve remarkable progress, yet they are typically trained at limited resolutions due to the lack of high-resolution data...
arXiv:2607. 06982v1 Announce Type: cross Abstract: Convolutional neural networks (CNNs) have demonstrated encouraging results in image classification tasks.
The paper introduces Gaussian-Grid, a hybrid neural representation that blends structured feature grids with unstructured 2D Gaussian distributions. It dynamically allocates computation based on local frequency content, enabling efficient image representation. Experiments on high‑resolution natural images show significant gains in decoding speed and PCIe data traffic reduction compared to JPEG + ZIP, while preserving small memory footprints and high reconstruction fidelity.