A new semi-supervised audio analysis model leveraging bi-path feature extraction and attention mechanisms boosts sound event detection accuracy by 12.7% (PSDS1 0.489, PSDS2 0.771), enabling smarter smart-home, security, wildlife and industrial monitoring.

A research team leverages chlorophyll fluorescence (ChlF) imaging, a cutting-edge technique, to identify reliable pre-symptomatic diagnostic indicators for rice blast and brown spot.

A research team has developed an innovative remote sensing model that significantly improves the accuracy of monitoring nitrogen accumulation in wheat crops.

A research team now proposes a novel Multi-Granularity Alignment (MGA) domain adaptation framework that dramatically improves cross-domain detection accuracy, enabling deep learning models to maintain high performance across datasets collected in different countries and conditions.