An involution-infused DenseNet student, compressed via knowledge distillation and weight pruning, achieves 96.99 to 98.63 percent accuracy with about 0.29 million parameters on two plant disease datasets.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Involution-Infused DenseNet with Two-Step Compression for Resource-Efficient Plant Disease Classification
An involution-infused DenseNet student, compressed via knowledge distillation and weight pruning, achieves 96.99 to 98.63 percent accuracy with about 0.29 million parameters on two plant disease datasets.