EUVIMEDEUVIMED entdecken
Crossref

Hybrid architecture of transformer and convolutional neural network with a multi-scale deformable attention mechanism for semantic segmentation task

R.R. Otyrba, A.A. Sirota

Computer Optics · 01.02.2026

DOI: 10.18287/coj1686 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Computer Optics, online erschienen am 01.02.2026. Verfasst von R.R. Otyrba und 1 weiteren Autor:innen. Die Arbeit nennt 2 Literaturangaben. Gefördert von: Ministry of Science and Higher Education of the Russian Federation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

A hybrid neural network architecture, SegTwice, is proposed for the semantic segmentation task. It combines the strengths of transformers and convolutional neural networks within a unified encoder-decoder framework. The original architecture of the encoding network, TWICE-DA, is presented, featuring a hierarchical structure with four levels. New architectural solutions are introduced and justified within the transformer blocks, which differ from known analogs: a multi-scale perception unit, a channel attention module, a deformable attention module, and a convolutional feedforward network module. Experiments on image classification tasks were conducted to assess the feature extraction effectiveness of TWICE-DA on datasets of varying complexity. It is shown that TWICE-DA demonstrates high quality, outperforming most modern models in terms of accuracy and computational complexity. The integration of TWICE-DA into the semantic segmentation network structure is achieved by adding a lightweight MLP decoder, ultimately realizing the SegTwice architecture. Experiments conducted on standard aerospace datasets, LoveDA and Potsdam, revealed that the proposed SegTwice network demonstrates competitive performance, matching traditional models and modern transformers in accuracy, and in some cases, outperforming them. Notably, SegTwice was trained "from scratch" without pre-training on large datasets, highlighting its resilience to overfitting in scenarios with limited data.

Förderung

  • Ministry of Science and Higher Education of the Russian Federation

Zitieren

APA

Otyrba, R., & Sirota, A. (2026). Hybrid architecture of transformer and convolutional neural network with a multi-scale deformable attention mechanism for semantic segmentation task. Computer Optics. https://doi.org/10.18287/coj1686

Vancouver

Otyrba R, Sirota A. Hybrid architecture of transformer and convolutional neural network with a multi-scale deformable attention mechanism for semantic segmentation task. Computer Optics. 2026. doi:10.18287/coj1686

BibTeX
@article{Otyrba2026,
  author = {Otyrba, R.R. and Sirota, A.A.},
  title = {Hybrid architecture of transformer and convolutional neural network with a multi-scale deformable attention mechanism for semantic segmentation task},
  journal = {Computer Optics},
  year = {2026},
  doi = {10.18287/coj1686}
}
Kostenlos starten →