Skip to main content

AINTLab Research Team Publishes Breakthrough in Agricultural AI

· One min read


한국어로 읽기: http://pr.sejong.ac.kr/sejongnewspaperview.do?boardType=6&pkid=67698


The AINTLab research team, led by Dr. Syafrudin, has published a groundbreaking study in the prestigious journal Computers and Electronics in Agriculture (impact factor: 7.7, top 1.7% in its field). The paper introduces FLTrans-Net (Transformer-based Feature Learning Network), an innovative algorithm for accurately detecting wheat ears in challenging agricultural environments.

Traditional convolutional neural networks (CNNs) struggle with small or overlapping wheat ears, but FLTrans-Net addresses this with a multi-scale fusion block, transformer encoder, spatial attention block, and a lightweight RetinaNet detection block. The algorithm enhances detection accuracy and is optimized for resource-constrained devices, promising to improve agricultural efficiency significantly.

Dr. Syafrudin, a faculty member at Sejong University since 2022 and recognized in the 2024 Global Top 2% Scientists list (Stanford/Elsevier), expressed gratitude to collaborators from AINTLab, HITEC University, and Kaunas University of Technology. He stated, “FLTrans-Net represents a significant step forward in applying AI to real-world agricultural challenges. We will continue to innovate in this area.”

Read the full study here: DOI: 10.1016/j.compag.2024.109706.

You may request the full text for personal use via https://www.researchgate.net/publication/386742803_FLTrans-Net_Transformer-based_feature_learning_network_for_wheat_head_detection