Leaf-DETR Sharpens Dense Leaf Detection in Crowded Kiwifruit Canopies
· newswise.com · popularity 19/100 · 0 similar stories today · Physics & Chemistry (Science)
Read the original at newswise.com →
Summary
<img src="https://www.newswise.com/legacy/image.php?image=https://media.newswise.com/pio/articles/858175/6af06199a6db6_28_858175.jpg&width=100&height=150" alt="Newswise image" />Accurate leaf monitoring is essential for smart agriculture because leaves reveal photosynthesis, growth, stress, and disease status. Dense leaf detection is especially difficult: overlapping leaves, similar textures, and heavy occlusion cause missed detections and unstable training. To address this, researchers developed Leaf-DETR, an end-to-end detection transformer (DETR) framework with a Progressive Feature Fusion Pyramid Network (P-FPN) and a Crowded Query Refinement (CQR) strategy. They also built a large dense kiwifruit leaf dataset containing 1,696 images and 85,375 annotated boxes. On this benchmark, Leaf-DETR improved mean average precision at intersection-over-union 0.50 (mAP@50) by 1% and average recall at 300 detections (AR@300) by 1.4% over its baseline, while converging quickly and generalizing to surveillance images and other crops.
Found in: Newswise. Rating basis: single outlet; outlet tier 2 (unlisted); 1 feeds; feed position 9.