Fruit-HSNet, using Fourier magnitude features plus the central pixel spectrum with learned fusion, reports 70.73% accuracy on DeepHS Fruit, a 12 point gain over the prior SOTA.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fruit-HSNet: A Machine Learning Approach for Hyperspectral Image-Based Fruit Ripeness Prediction
Fruit-HSNet, using Fourier magnitude features plus the central pixel spectrum with learned fusion, reports 70.73% accuracy on DeepHS Fruit, a 12 point gain over the prior SOTA.