UnfoldCRF treats refinement as inference in a conditional random field over pixel labels and latent region variables, and zeroing individual messages shows where the corrections come from: the pairwise messages mostly fix boundaries, the region messages mostly fix non-boundary errors.
Abstract
Learned mask refiners improve segmentation accuracy, but it is hard to tell how much of the improvement comes from explicit structure rather than from extra capacity, and whether it holds up when the mask generator or its error distribution changes. UnfoldCRF treats refinement as inference in a conditional random field over pixel labels and latent region variables. Its energy has a corrected unary term, learned local pairwise interactions, and image-conditioned latent-region consistency, with a null state that lets a region with weak label agreement withdraw from the consistency term; inference unrolls damped mean-field updates on this one energy. To isolate the effect of structure, we compare against recurrent black-box refiners that read the same inputs and receive the same parameter budget, stage count, and supervision. On COD10K, UnfoldCRF beats the strongest matched control by 1.0 $F^\omega_\beta$ point, improves all four COD metrics, and lowers the fraction of images made worse from 11.7\% to 8.5\%. Under a train-once protocol over five datasets and several mask sources, the 2.6M-parameter variant gains 4.2 mean $\Delta$IoU against 2.0 for its control, and a variant built on frozen DINOv2 features matches the strongest foundation-model refiner with about a seventh of its resident parameters while staying ahead of its own control. On mask generators never seen in training, the gain is 2.0 $F^\omega_\beta$ points against 0.6 for the control. Zeroing individual messages shows where the corrections come from: the pairwise messages mostly fix boundaries, the region messages mostly fix non-boundary errors. Code and supporting materials will be publicly released.
PredErase, a training-free inference procedure on frozen FLUX.2 and I-JEPA, improves the native FLUX.2 backbone and is supported is training-free object-and-effect editing of frozen Fill, not replacement of paired-data erasers.
Wai-Kit Xiu, Qiang Lu, Jun-Biao Chen et al.· 0 citations
Visual classifiers are expected to generalize under data shifts, target shifts, and their combinations, yet most existing methods focus on domain invariance while failing to address intra-image predictive sufficiency. We investigate the structural hypothesis that each image contains a sample-adaptive oracle intra-image...
A deep ensemble of the two co-first-place 2025 models is formed and a lightweight residual refiner trained on the ensemble's own outputs under an $\ell_1$ loss augmented with a structural-similarity term whose weight $\lambda$ the authors vary, which improves an already strong ensemble without any large-scale retrainin...
JAx (Just Align x), a prediction-supervision method that aligns clean-image predictions across noise levels, shows that prediction-space supervision provides a simple and principled alternative to representation alignment for pixel-space generative models.
Yu-Yao Zhang, Yu-Wei Hu, Zi-Yang Mai et al.· 0 citations
Text-guided image editing using rectified flow models such as Multimodal Diffusion Transformer (DiT) has demonstrated impressive generation quality. However, existing methods apply edits globally, inevitably modifying background regions unrelated to the intended semantic change. Our key observation is that the Text-to-...
Trong-Tai Dam Vu, Vinh-Tiep Nguyen· International Conference on...· 0 citations
Inspired by CRF potentials, a framework that disentangles soft pseudo-labels as unary supervision and binary edge maps as pairwise supervision is proposed that sets a new state-of-the-art of 56.5% mIoU on MS COCO among training-based methods.
Joshua Li, Yuri Boykov· arXiv.org· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.