grain-detection

1 article

Tiny AI model spots grain defects in maize, wheat, and rice with high accuracy

OpenAlex · 2026-07-10

Researchers built a compact AI model called LGC-Net that can quickly identify defects in five major grain crops using images, achieving nearly 90% accuracy while being small enough to run on low-power devices in the field.

LGC-Net-XXS achieves 89.59% average classification accuracy across five grain types with more than ten defect categories

The model uses only 0.93 million parameters and 0.43 billion floating-point operations, making it suitable for low-power edge devices

A new Fast Channel Additive Attention mechanism reduces computational complexity from quadratic to linear while retaining global pattern recognition

Species in Topic

Rice

Wheat

Maize

Barley

Sorghum

grain-detection

Species in Topic

Related Topics