◆ Computers and Electronics in Agriculture2026-05-28· Computer science
A multi-criteria evaluation framework for mobile weed classification: Balancing accuracy and efficiency across hardware tiers
Leonardo G. Sanches, Henrique Y. Shishido, Claiton de Oliveira, Danilo S. Sanches, Lucas D.H. Sampaio, Willian M. Watanabe, Silvio R. R. Sanches
原始摘要（英文原文）· Original abstract
• Framework evaluates DL models for weed classification. • MobileNetV3-Large and Small are Pareto-optimal in accuracy/efficiency. • Lightweight models outperform heavyweights in latency and memory. • Multi-tier benchmarking protocol ensures reproducibility. • Pareto-frontier analysis guides hardware-aware architecture selection.