AlGaN-Based UV LEDs With High Wall-Plug Efficiency

Wenting Gong

doi:10.1109/led.2026.3698638

Bloom–Weave–Balance: A sample-efficient framework for hyperspectral image classification

2026-08-01

0

PRE

AI

刘

刘荣 (Rong Liu)

J

Jiayang Huang

J

Jiaqi Yang *

DOI：10.1016/j.patcog.2026.114577

Original

Original request for help

Save

Abstract

En 中文

• Bloom expands data via target-centered, spectrally and spatially consistent pseudo-samples. • Weave integrates channel attention and multi-scale self-attention across scales. • Balance combines cross-entropy, focal, and Dice losses to reduce class imbalance. • BWB improves data use, feature learning, and class fairness in accurate HSI classification.

Keywords:

Hyperspectral image classification

Convolutional neural network

Transformer

Sample scarcity

Sample augmentation

Bloom–Weave–Balance: A sample-efficient framework for hyperspectral image classification

Abstract

Journal

Organization

Cited Papers

Citing Papers