New research explores diffusion model advancements for image and video generation · 9 sources tracked
ByPulseAugur Editorial·[16 sources]·
Multiple research papers released on arXiv explore advancements in diffusion models for image and video generation. These studies introduce novel techniques such as landmark-constrained acceleration for vector diffusion maps, tensor-based functionals for image enhancement, and channel representations for non-linear diffusion filtering. Other papers focus on improving the efficiency and generalization of diffusion models, including methods for fast sampling, elastic token compression, and waypoint diffusion transformers for pixel-space generation. Additionally, research addresses test-time scaling for video diffusion models through candidate recycling and discrete diffusion bridges for spatiotemporally aligned image translation.
AI
IMPACT
These papers introduce novel techniques and optimizations for diffusion models, potentially improving image and video generation quality and efficiency.
RANK_REASON
Multiple arXiv papers detailing novel research in diffusion models.
arXiv:2609.02016v1 Announce Type: cross Abstract: Image super-resolution, which aims to reconstruct high-resolution images from their low-resolution observations, is fundamental to medical imaging, remote sensing, surveillance, microscopy, and scientific visualization. Traditiona…
arXiv:2609.02160v1 Announce Type: cross Abstract: Diffusion models achieve high sample quality but remain expensive at inference time because sampling requires many sequential neural function evaluations (NFEs). Existing acceleration methods either use fixed step-skipping schedul…
arXiv cs.AI
TIER_1English(EN)·Zhengming Yu, Li Ma, Mingming He, Leo Isikdogan, Yuancheng Xu, Dmitriy Smirnov, Pablo Salamanca, Dao Mi, Pablo Delgado, Ning Yu, Julien Philip, Xin Li, Wenping Wang, Paul Debevec·
arXiv:2604.06161v3 Announce Type: replace-cross Abstract: Most digital videos are stored in 8-bit low dynamic range (LDR) formats, where much of the original high dynamic range (HDR) scene radiance is lost due to saturation and quantization. This loss of highlight and shadow deta…
Pixel-space diffusion has recently emerged as a promising direction for high-fidelity image generation by modeling images directly in the original pixel domain. However, pixel-space diffusion is computationally demanding due to the extremely high dimensionality of natural images.…
arXiv:2603.21247v2 Announce Type: replace-cross Abstract: We propose a landmark-constrained algorithm, LA-VDM (Landmark Accelerated Vector Diffusion Maps), to accelerate the Vector Diffusion Maps (VDM) framework built upon the Graph Connection Laplacian (GCL), which captures pair…
arXiv:2609.02036v1 Announce Type: new Abstract: Few-step diffusion models substantially compress temporal computation, making the spatial cost of each model evaluation an increasingly dominant source of inference latency. Progressive-resolution inference reduces this cost by perf…
arXiv:2609.00798v1 Announce Type: new Abstract: Pixel-space diffusion has recently emerged as a promising direction for high-fidelity image generation by modeling images directly in the original pixel domain. However, pixel-space diffusion is computationally demanding due to the …
arXiv:2609.00862v1 Announce Type: new Abstract: We show that a frozen generic text-to-image diffusion model can perform conditional inpainting across three evaluated natural-image domains with one fixed controller configuration, without inpainting-specific weight training, datase…
arXiv:2603.15132v3 Announce Type: replace Abstract: While recent Flow Matching models avoid the reconstruction bottlenecks of latent autoencoders by operating directly in pixel space, the raw pixel manifold provides little explicit semantic organization, making target-specific tr…
arXiv:2608.29322v1 Announce Type: new Abstract: Recent video diffusion models have achieved remarkable generation quality, but high-fidelity results still largely depend on closed-source systems or costly large-scale infrastructure. Test-time scaling (TTS) offers a training-free …
arXiv cs.CV
TIER_1English(EN)·Eduard Zamfir, Christian Reisswig, Zongwei Wu, Yongqin Xian, Radu Timofte·
arXiv:2608.29281v1 Announce Type: new Abstract: Natural images concentrate their detail in a small fraction of the frame, yet diffusion models spend a full token on every patch, in every layer and at every timestep. The waste is largest in pixel-space models, with no autoencoder …
arXiv:2608.29233v1 Announce Type: new Abstract: We present the winning solution to the ACM Multimedia 2026 Grand Challenge on Single-Image Guided Multi-Angle Image Synthesis. It ranks first among 293 registered teams; 56 teams obtained at least one scored submission on the public…
arXiv cs.CV
TIER_1English(EN)·Christian Heinemann, Freddie {\AA}str\"om, George Baravdish, Kai Krajsek, Michael Felsberg, Hanno Scharr·
arXiv:2608.29227v1 Announce Type: new Abstract: In this work we propose a novel non-linear diffusion filtering approach for images based on their channel representation. To derive the diffusion update scheme we formulate a novel energy functional using a soft-histogram representa…
arXiv:2608.29997v1 Announce Type: new Abstract: We propose Discrete Diffusion Bridges (DDB), a novel framework designed to resolve the fundamental spatiotemporal misalignment of standard discrete diffusion in image translation and generation. By corrupting data into a pure mask s…
arXiv cs.CV
TIER_1English(EN)·Freddie {\AA}str\"om, Michael Felsberg, George Baravdish·
arXiv:2608.29164v1 Announce Type: new Abstract: In this work, we introduce a novel tensor-based functional for targeted image enhancement and denoising. Via explicit regularization, our formulation incorporates application dependent and contextual information using first principl…
arXiv cs.CV
TIER_1English(EN)·Zhenyu Zhou, Defang Chen, Siwei Lyu, Chun Chen, Can Wang·
arXiv:2603.00763v2 Announce Type: replace Abstract: Text-to-image diffusion models have achieved unprecedented success but still struggle to produce high-quality results under limited sampling budgets. Existing training-free sampling acceleration methods are typically developed i…
