S²-FracMix creates label-preserving augmented samples via intra-image self-saliency patch mixing and fractal pattern injection, claiming SOTA results across seven benchmarks in classification, robustness, detection, and transfer tasks.
author Marcu, A
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
representative citing papers
PointCaM proposes a cut-and-mix mechanism with an Unknown-Point Simulator and Estimator to improve open-set recognition on point clouds by simulating out-of-distribution data and using multi-level features.
On vein-recognition benchmarks, mixup-style augmentations win on clean accuracy but hurt calibration and adversarial robustness, while simple geometric transforms usually hurt performance.
LHSD estimates local intrinsic dimension in high-D spaces by spectral filtering of the log-density Hessian via SLQ to isolate zero-curvature tangent directions.
Nonlinear transformations enable DNNs to achieve substantial test accuracy gains (0.34% to 249.59%) on unlearnable CIFAR10 datasets from twelve protection methods, outperforming a recent linear baseline.
citing papers explorer
-
$S^{2}$-FracMix: Label-Preserving Self-Saliency Mixup Augmentation
S²-FracMix creates label-preserving augmented samples via intra-image self-saliency patch mixing and fractal pattern injection, claiming SOTA results across seven benchmarks in classification, robustness, detection, and transfer tasks.
-
PointCaM: Cut-and-Mix for Open-Set Point Cloud Learning
PointCaM proposes a cut-and-mix mechanism with an Unknown-Point Simulator and Estimator to improve open-set recognition on point clouds by simulating out-of-distribution data and using multi-level features.
-
AGVBench: A Reliability-Oriented Benchmark of Data Augmentation for Vein Recognition
On vein-recognition benchmarks, mixup-style augmentations win on clean accuracy but hurt calibration and adversarial robustness, while simple geometric transforms usually hurt performance.
-
Local Hessian Spectral Filtering for Robust Intrinsic Dimension Estimation
LHSD estimates local intrinsic dimension in high-D spaces by spectral filtering of the log-density Hessian via SLQ to isolate zero-curvature tangent directions.
-
Nonlinear Transformations Against Unlearnable Datasets
Nonlinear transformations enable DNNs to achieve substantial test accuracy gains (0.34% to 249.59%) on unlearnable CIFAR10 datasets from twelve protection methods, outperforming a recent linear baseline.