analiza aiqbench
This arXiv preprint proposes SIFR, a dual-stream vision model that uses frequency-domain techniques to fuse frozen CLIP and DINOv2 features and stabilize decoder features, aiming to make semantic segmentation hold up on unseen domains such as adverse weather. The reported gains are relative to prior methods on several benchmarks, but the abstract gives no numbers, and the results come from a single unreviewed v1 preprint. Readers should check the benchmark tables and ablations before treating the approach as a practical improvement over existing foundation-model baselines.
Spectral Invariant Fusion and Refinement for Domain Generalized Semantic Segmentation
arXiv:2610.39926v1 Announce Type: cross Abstract: Domain generalized semantic segmentation is challenging because it requires accurate dense prediction on unseen domains without any access to target-domain data during training. Existing methods improve transferability through data diversification, invariant learning, or direct use of foundation models, but they often struggle to simultaneously suppress domain-sensitive interference and preserve fine-grained spatial cues. In this work, we propose SIFR, a frequency-aware dual-stream framework that tackles domain shift at both the fusion and decoding stages. SIFR uses spectral modeling as a unified principle to regulate both heterogeneous prior interaction and dense decoder refinement. Instead of directly fusing frozen CLIP and DINOv2 features, SIFR performs spectral-guided invariant fusion to filter out domain-sensitive responses while preserving complementary semantic and structural information. In addition, SIFR incorporates a spectral decoder refinement module that stabilizes multi-scale decoder features through low-frequency statistics-based normalization and adaptive frequency selection, thereby reducing the propagation of residual domain bias in dense prediction. This design enables SIFR to jointly improve domain robustness and segmentation precision under severe distribution shifts. Extensive experiments on multiple DGSS benchmarks, including synthetic-to-real, real-to-real, and adverse-weather settings, demonstrate consistent improvements over previous state-of-the-art methods.
Abstract:Domain generalized semantic segmentation is challenging because it requires accurate dense prediction on unseen domains without any access to target-domain data during training. Existing methods improve transferability through data diversification, invariant learning, or direct use of foundation models, but they often struggle to simultaneously suppress domain-sensitive interference and preserve fine-grained spatial cues. In this work, we propose SIFR, a frequency-aware dual-stream framework that tackles domain shift at both the fusion and decoding stages. SIFR uses spectral modeling as a unified principle to regulate both heterogeneous prior interaction and dense decoder refinement. Instead of directly fusing frozen CLIP and DINOv2 features, SIFR performs spectral-guided invariant fusion to filter out domain-sensitive responses while preserving complementary semantic and structural information. In addition, SIFR incorporates a spectral decoder refinement module that stabilizes multi-scale decoder features through low-frequency statistics-based normalization and adaptive frequency selection, thereby reducing the propagation of residual domain bias in dense prediction. This design enables SIFR to jointly improve domain robustness and segmentation precision under severe distribution shifts. Extensive experiments on multiple DGSS benchmarks, including synthetic-to-real, real-to-real, and adverse-weather settings, demonstrate consistent improvements over previous state-of-the-art methods.
Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
arXiv:2610.39926 [cs.CV]
(or arXiv:2610.39926v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.39926 Focus to learn more arXiv-issued DOI via DataCite
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 · 283 cuvinte
Subiecte conexe
Uncertainty Gates the Gradient: Adaptive Updates for Generalized Few-Shot 3D Segmentation
arXiv · 21 oct. 2026
Stable Continuous-Time Consistency Distillation: An Empirical Study with a Multistep Extension
arXiv · 27 oct. 2026
ROBUST-WT: Robust Uncertainty-aware Segmentation Transform via Whitening and Training Enhancements
arXiv · 3 iun. 2026
How Frontier AI Labs Acquire Visual Training Data for Multimodal Models
Roboflow Blog · 2 nov. 2026