Systematic analytic-geometric augmentation for robust deep learning

ТАТУ хабарлари · 2024-yil

Annotatsiya

Robustness to geometric distortions and distribution shifts remains a major challenge in deep learning.Standard data augmentation methods, such as random rotations and translations, improve generalization but often fail to systematically cover the geometric transformation space.In this work, we propose an analytic-geometric data augmentation framework that employs controlled affine transformations-including fixed-angle rotations, shearing, and scalingdefined analytically for each training sample.We hypothesize that such systematic augmentation better exposes neural networks to the full spectrum of geometric variations, enhancing their ability to generalize and remain robust under previously unseen conditions.We conduct extensive experiments on MNIST, FashionMNIST, and CIFAR-10 (grayscale), training lightweight convolutional models with both standard and analytic-geometric augmentations.Evaluation on clean, noisy, and rotated test sets demonstrates that analytic-geometric augmentation consistently yields higher accuracy and significantly improved robustness across all scenarios.Notably, our method boosts performance under noise and out-of-distribution rotations by up to 30% compared to conventional augmentations.Our findings highlight the value of explicitly structured augmentation for building more robust and reliable computer vision systems, and suggest that analytic-geometric methods should be incorporated into standard deep learning pipelines for applications where robustness is critical.

Maqola ma’lumotlari
JurnalТАТУ хабарлари
NashriyotТошкент ахборот технологиялари университети
ISSN2010-9857, 2181-1083
Nashr sanasi2024-12-14
TilIngliz
Ilmiy sohaTexnika
DOI10.61663/243stuitmct2
Indeksga qo‘shilgan2026-09-02

DOI: 10.61663/243stuitmct2 · Maqolaning asl sahifasi

Ilmiy soha

Texnika

Iqtibos keltirish

APANoma’lum muallif (2024). SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING. ТАТУ хабарлари. https://doi.org/10.61663/243stuitmct2
GOST (OAK)Noma’lum muallif SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING // ТАТУ хабарлари. – 2024. – DOI: 10.61663/243stuitmct2.
MLANoma’lum muallif. “SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING.” ТАТУ хабарлари, 2024. https://doi.org/10.61663/243stuitmct2
ChicagoNoma’lum muallif. “SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING.” ТАТУ хабарлари (2024). https://doi.org/10.61663/243stuitmct2
HarvardNoma’lum muallif (2024) ‘SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING’, ТАТУ хабарлари. https://doi.org/10.61663/243stuitmct2
BibTeX
@article{article2024,
  title = {SYSTEMATIC ANALYTIC-GEOMETRIC AUGMENTATION FOR ROBUST DEEP LEARNING},
  author = {Unknown},
  journal = {ТАТУ хабарлари},
  year = {2024},
  doi = {10.61663/243stuitmct2},
  url = {https://doi.org/10.61663/243stuitmct2}
}

Boshqa maqola uchun iqtibos tayyorlash

ТАТУ хабарлари jurnalidan boshqa maqolalar

ТАТУ хабарлари — barcha maqolalar · 2024-yil arxivi

Texnika sohasidagi boshqa maqolalar

Texnika sohasidagi OAK jurnallari