Large-scale text-to-image models are attractive backbones for dense prediction because RGB generation pretraining learns rich semantic, structural, and geometric priors. Existing generative and editing approaches reuse these priors by casting dense prediction as target generation: annotations such as depth, normals, alpha mattes, masks, and heatmaps are encoded into an RGB-trained VAE latent space and decoded back as image-like targets.
DAGS introduces a lightweight, attention‑free conditioning scheme that disentangles appearance and geometry for a frozen image diffusion transformer (DiT), enabling high‑fidelity, temporally stable renders with independent control. Two small convolutional encoders generate per‑frame conditioning features, which are injected as learned residuals into the image tokens, avoiding the quadratic cost of attention. Coupled with a recurrent lighting stabilizer and a training‑free temporal guidance term, DAGS transforms a per‑frame image model into a streaming renderer that outperforms real‑time denoisers and diffusion renderers in PSNR and temporal stability while requiring far less compute than path tracing.
By Karthik Mohan Kumar, Damian Andrysiak, Pedro Antonio Pena, Kunal Tyagi, Rama Harihara
arXiv:2609.22789v1 Announce Type: new
Abstract: Image-to-layer decomposition converts a flattened image into editable RGBA layers, enabling element-level editing in design workflows. Existing diffusi...
By Zelin Jia, Zhao Zhang, Zhicong Tang, Yuhui Yuan, Shixia Liu
PixelUMM is an encoder‑free model that unifies image and video understanding and generation directly in pixel space. It represents images as spatial patches and videos as spatiotemporal tubelets, feeding both through single‑layer linear projections into a shared multimodal backbone. The Mixture‑of‑Transformers architecture blends shared attention with task‑specific parameters, enabling autoregressive text prediction, pixel‑space flow matching, and clean‑pixel video generation, and experiments show competitive performance across tasks while providing design insights for future pixel‑space multimodal models.
By Cong Wei, Xuanchi Ren, Bryan Chu, Weiming Ren, Huan Ling, Jiahui Huang, Laura Leal-Taix\'e, Sanja Fidler, Wenhu Chen, Zian Wang, Jay Zhangjie Wu
JointLight introduces an asynchronous joint-modeling approach for human-centric video relighting, enabling fine-grained illumination control by jointly denoising RGB and irradiance frames during two-stage training. The method employs a domain matching strategy to bridge gaps between synthetic irradiance cues and real irradiance maps, and uses an asynchronous hijacked generation pipeline at inference. A newly created multi-illumination synthetic video dataset with paired irradiance maps supports training, and the model demonstrates strong generalization to real-world scenarios, outperforming prior methods in lighting fidelity and generative quality.
By Henglei Lv, Bailin Deng, Xiaoqiang Liu, Yichen Wang, Haoxian Zhang, Pengfei Wan, Lin Gao
arXiv:2606.13345v2 Announce Type: replace
Abstract: Existing 3D scene editing methods typically rely on per-scene optimization over explicit 3D representations or cascaded edit-and-reconstruct pipeli...
By Xinnan Zhu, Ruijie Xu, Jiayu Ying, Daoguo Dong, Jiachen Xu, Yuan Xie, Xin Tan
arXiv:2606.21030v2 Announce Type: replace-cross
Abstract: Diffusion-based image compression has achieved strong perceptual quality at ultra-low bitrates. However, existing codecs are often tied to sp...
By Yinhuan Huang, Hao Cao, Pu chen, Wenqi Guo, Jungong Han, Zhijin Qin
arXiv:2610.09125v1 Announce Type: new
Abstract: Reference-based object compositing inserts or replaces an object using a background image, a reference image, and a 2D compositing mask. These inputs g...
By Sanghyun Jo, Chae Yeon Lim, Donghwan Lee, Sihyun Kim, Soo Ye Kim, Kyungsu Kim
arXiv:2610.09450v1 Announce Type: new
Abstract: Pixel-space diffusion models avoid the lossy VAE of latent models, which suggests an advantage on downstream tasks where fine-grained detail matters. W...
By Hanqiu Li Cai (SperidLabs), Chema Garabito (SperidLabs)
IDM-Net is a lightweight Illumination-Decoupled Modulation Network designed for low-light image enhancement. It uses a dual-encoder architecture: a structure encoder extracts multi-scale appearance features from the RGB image, while a lightweight illumination encoder learns illumination priors from the decoupled luminance channel. An Illumination-Guided Modulation module injects these priors into the decoder via spatially adaptive affine modulation, and a Feature Refinement Block progressively suppresses artifacts and recovers fine details, achieving competitive performance on standard benchmarks with good balance between quality and efficiency.
By Cheng-Yen Hsiao, Jing-Ming Guo
arXiv:2610.35515v1 Announce Type: new
Abstract: Diffusion-based real-world image super-resolution suffers from limited reconstruction fidelity due to information loss in VAE-based latent representati...
By Wenjie Li, Du Chen, Hui Wang, Linlong Fan, Youwei Lv, Zeyi Chen, Zehan Chen, Jinwei Chen, Heng Guo, Zhanyu Ma, Qingnan Fan
4D generation synthesizes dynamic 3D scenes from conditions such as text or images. Existing methods either reconstruct generated RGB videos with a separate 4D model or adapt a particular video generator to predict geometry directly.